EP2812894A2 - Motion picture project management system - Google Patents
Motion picture project management systemInfo
- Publication number
- EP2812894A2 EP2812894A2 EP13746718.9A EP13746718A EP2812894A2 EP 2812894 A2 EP2812894 A2 EP 2812894A2 EP 13746718 A EP13746718 A EP 13746718A EP 2812894 A2 EP2812894 A2 EP 2812894A2
- Authority
- EP
- European Patent Office
- Prior art keywords
- project
- frame
- mask
- shot
- computer
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Withdrawn
Links
- 230000033001 locomotion Effects 0.000 title claims abstract description 232
- 239000002131 composite material Substances 0.000 claims description 86
- 230000000873 masking effect Effects 0.000 claims description 56
- 238000007726 management method Methods 0.000 claims description 43
- 238000004519 manufacturing process Methods 0.000 claims description 40
- 238000012384 transportation and delivery Methods 0.000 claims description 15
- 238000009432 framing Methods 0.000 claims 1
- 238000012946 outsourcing Methods 0.000 claims 1
- 238000012545 processing Methods 0.000 abstract description 25
- 238000003860 storage Methods 0.000 abstract description 14
- 238000010191 image analysis Methods 0.000 abstract description 2
- 238000000034 method Methods 0.000 description 124
- 238000013519 translation Methods 0.000 description 55
- 230000014616 translation Effects 0.000 description 55
- 238000009877 rendering Methods 0.000 description 53
- 238000006243 chemical reaction Methods 0.000 description 42
- 230000008569 process Effects 0.000 description 38
- 230000000007 visual effect Effects 0.000 description 34
- 239000013078 crystal Substances 0.000 description 33
- 238000013461 design Methods 0.000 description 33
- 210000003128 head Anatomy 0.000 description 33
- 238000012552 review Methods 0.000 description 32
- 239000003086 colorant Substances 0.000 description 24
- 230000003068 static effect Effects 0.000 description 23
- 238000000275 quality assurance Methods 0.000 description 22
- 239000000872 buffer Substances 0.000 description 20
- 239000013598 vector Substances 0.000 description 18
- 238000004364 calculation method Methods 0.000 description 17
- 239000000203 mixture Substances 0.000 description 15
- 230000000644 propagated effect Effects 0.000 description 14
- 230000006870 function Effects 0.000 description 13
- 230000003416 augmentation Effects 0.000 description 11
- 230000000694 effects Effects 0.000 description 11
- 230000009471 action Effects 0.000 description 10
- 238000004422 calculation algorithm Methods 0.000 description 10
- 238000006073 displacement reaction Methods 0.000 description 10
- 238000004458 analytical method Methods 0.000 description 9
- 238000012937 correction Methods 0.000 description 9
- 230000001934 delay Effects 0.000 description 9
- 238000003708 edge detection Methods 0.000 description 9
- 239000011521 glass Substances 0.000 description 9
- 238000002156 mixing Methods 0.000 description 9
- 238000012986 modification Methods 0.000 description 9
- 230000004048 modification Effects 0.000 description 9
- 238000005516 engineering process Methods 0.000 description 7
- 230000001965 increasing effect Effects 0.000 description 7
- 230000008859 change Effects 0.000 description 6
- 238000009826 distribution Methods 0.000 description 6
- 239000000428 dust Substances 0.000 description 6
- 230000037433 frameshift Effects 0.000 description 6
- 239000003973 paint Substances 0.000 description 6
- 238000010422 painting Methods 0.000 description 6
- 238000012360 testing method Methods 0.000 description 6
- 230000009466 transformation Effects 0.000 description 6
- 238000001514 detection method Methods 0.000 description 5
- 238000003909 pattern recognition Methods 0.000 description 5
- 230000007704 transition Effects 0.000 description 5
- 238000003491 array Methods 0.000 description 4
- 230000002902 bimodal effect Effects 0.000 description 4
- 230000015556 catabolic process Effects 0.000 description 4
- 238000000605 extraction Methods 0.000 description 4
- 230000008676 import Effects 0.000 description 4
- 235000012054 meals Nutrition 0.000 description 4
- 230000010363 phase shift Effects 0.000 description 4
- 230000009467 reduction Effects 0.000 description 4
- 230000006835 compression Effects 0.000 description 3
- 238000007906 compression Methods 0.000 description 3
- 239000000470 constituent Substances 0.000 description 3
- 230000008030 elimination Effects 0.000 description 3
- 238000003379 elimination reaction Methods 0.000 description 3
- 230000002708 enhancing effect Effects 0.000 description 3
- 238000013439 planning Methods 0.000 description 3
- 230000002829 reductive effect Effects 0.000 description 3
- 238000005070 sampling Methods 0.000 description 3
- 238000012935 Averaging Methods 0.000 description 2
- XEEYBQQBJWHFJM-UHFFFAOYSA-N Iron Chemical compound [Fe] XEEYBQQBJWHFJM-UHFFFAOYSA-N 0.000 description 2
- 238000013459 approach Methods 0.000 description 2
- 230000001680 brushing effect Effects 0.000 description 2
- 238000004891 communication Methods 0.000 description 2
- 238000010276 construction Methods 0.000 description 2
- 238000011156 evaluation Methods 0.000 description 2
- 239000000284 extract Substances 0.000 description 2
- 238000011049 filling Methods 0.000 description 2
- 230000002452 interceptive effect Effects 0.000 description 2
- 238000010409 ironing Methods 0.000 description 2
- 238000013507 mapping Methods 0.000 description 2
- 239000011159 matrix material Substances 0.000 description 2
- 238000005457 optimization Methods 0.000 description 2
- 230000001151 other effect Effects 0.000 description 2
- 230000000750 progressive effect Effects 0.000 description 2
- 238000011160 research Methods 0.000 description 2
- 230000000717 retained effect Effects 0.000 description 2
- 230000006641 stabilisation Effects 0.000 description 2
- 238000011105 stabilization Methods 0.000 description 2
- 230000009897 systematic effect Effects 0.000 description 2
- 238000012546 transfer Methods 0.000 description 2
- 238000000844 transformation Methods 0.000 description 2
- 238000011144 upstream manufacturing Methods 0.000 description 2
- 101100018566 Neurospora crassa (strain ATCC 24698 / 74-OR23-1A / CBS 708.71 / DSM 1257 / FGSC 987) tif-1 gene Proteins 0.000 description 1
- ZVQOOHYFBIDMTQ-UHFFFAOYSA-N [methyl(oxido){1-[6-(trifluoromethyl)pyridin-3-yl]ethyl}-lambda(6)-sulfanylidene]cyanamide Chemical compound N#CN=S(C)(=O)C(C)C1=CC=C(C(F)(F)F)N=C1 ZVQOOHYFBIDMTQ-UHFFFAOYSA-N 0.000 description 1
- 230000003466 anti-cipated effect Effects 0.000 description 1
- 230000000712 assembly Effects 0.000 description 1
- 238000000429 assembly Methods 0.000 description 1
- 230000003190 augmentative effect Effects 0.000 description 1
- 230000006399 behavior Effects 0.000 description 1
- 230000008901 benefit Effects 0.000 description 1
- 238000003339 best practice Methods 0.000 description 1
- 239000004568 cement Substances 0.000 description 1
- 238000010425 computer drawing Methods 0.000 description 1
- 238000013500 data storage Methods 0.000 description 1
- 230000007123 defense Effects 0.000 description 1
- 238000002716 delivery method Methods 0.000 description 1
- 238000011161 development Methods 0.000 description 1
- 230000018109 developmental process Effects 0.000 description 1
- 238000002224 dissection Methods 0.000 description 1
- 230000007717 exclusion Effects 0.000 description 1
- 230000006872 improvement Effects 0.000 description 1
- 230000000977 initiatory effect Effects 0.000 description 1
- 230000010354 integration Effects 0.000 description 1
- 229910052742 iron Inorganic materials 0.000 description 1
- PMHURSZHKKJGBM-UHFFFAOYSA-N isoxaben Chemical compound O1N=C(C(C)(CC)CC)C=C1NC(=O)C1=C(OC)C=CC=C1OC PMHURSZHKKJGBM-UHFFFAOYSA-N 0.000 description 1
- 238000011068 loading method Methods 0.000 description 1
- 238000010606 normalization Methods 0.000 description 1
- 230000003287 optical effect Effects 0.000 description 1
- 238000004091 panning Methods 0.000 description 1
- 239000002245 particle Substances 0.000 description 1
- 238000002360 preparation method Methods 0.000 description 1
- 238000003672 processing method Methods 0.000 description 1
- 238000003908 quality control method Methods 0.000 description 1
- 238000007670 refining Methods 0.000 description 1
- 230000008929 regeneration Effects 0.000 description 1
- 238000011069 regeneration method Methods 0.000 description 1
- 230000002441 reversible effect Effects 0.000 description 1
- 238000001228 spectrum Methods 0.000 description 1
- 230000007480 spreading Effects 0.000 description 1
- 238000003892 spreading Methods 0.000 description 1
- 238000012549 training Methods 0.000 description 1
- 238000009941 weaving Methods 0.000 description 1
Classifications
-
- G—PHYSICS
- G11—INFORMATION STORAGE
- G11B—INFORMATION STORAGE BASED ON RELATIVE MOVEMENT BETWEEN RECORD CARRIER AND TRANSDUCER
- G11B27/00—Editing; Indexing; Addressing; Timing or synchronising; Monitoring; Measuring tape travel
- G11B27/10—Indexing; Addressing; Timing or synchronising; Measuring tape travel
- G11B27/34—Indicating arrangements
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q10/00—Administration; Management
- G06Q10/06—Resources, workflows, human or project management; Enterprise or organisation planning; Enterprise or organisation modelling
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q10/00—Administration; Management
- G06Q10/10—Office automation; Time management
- G06Q10/103—Workflow collaboration or project management
Definitions
- Patent 7,181,081, filed June 18 th , 2003 which is a national stage entry of Patent Cooperation Treaty Application Serial No. PCT/US02/14192, filed May 6 th , 2003, which claims the benefit of United States Provisional Patent Application 60/288,929, filed May 4 th , 2001, the specifications of which are all hereby incorporated herein by reference.
- One or more embodiments of the invention are related to the field of project management in the motion picture industry and relates to reviewers, production managers who manage artists. Production managers are also known as "production" for short. Artists utilize image analysis and image enhancement and computer graphics processing for example to convert two-dimensional images into three-dimensional images associated with a motion picture or otherwise create or alter motion pictures. More particularly, but not by way of limitation, one or more embodiments of the invention enable motion picture project management system configured to efficiently manage projects related to motion pictures to manage assets, control costs, predict budgets and profit margins, reduce archival storage and otherwise provide displays tailored to specific roles to increase worker efficiency.
- the region is masked by an operator via a one -bit painted overlay and operator manipulated using a digital paintbrush method frame by frame to match the movement.
- each region is outlined or rotoscoped by an operator using vector polygons, which are then adjusted frame by frame by the operator, to create animated masked ROIs.
- Various masking technologies are generally also utilized in the conversion of 2D movies to 3D movies.
- the color transform lookup tables and regions selected are applied and modified manually to each frame in succession to compensate for changes in the image data that the operator detects visually. All changes and movement of the underlying luminance gray scale is subjectively detected by the operator and the masks are sequentially corrected manually by the use of an interface device such as a mouse for moving or adjusting mask shapes to compensate for the detected movement.
- the underlying gray scale is a passive recipient of the mask containing pre-selected color transforms with all modifications of the mask under operator detection and modification.
- the mask information does not contain any information specific to the underlying luminance gray scale and therefore no automatic position and shape correction of the mask to correspond with image feature displacement and distortion from one frame to another is possible.
- the iterative workflow includes masking objects in each frame, adding depth and then rendering the frame into left and right viewpoints forming an anaglyph image or a left and right image pair.
- the typical workflow involves an "iteration”, i.e., sending the frames back to the workgroup responsible for masking the objects, (which can be in a country with cheap unskilled labor half way around the world), after which the masks are sent to the workgroup responsible for rendering the images, (again potentially in another country), after which the rendered image pair is sent back to the quality assurance group.
- iterations of a complicated frame This is known as “throw it over the fence" workflow since different workgroups work independently to minimize their current work load and not as a team with overall efficiency in mind.
- existing motion picture projects sometimes require three people to review an edited frame in some cases, e.g., a person to locate the resource amongst a large number of resources, a person to review the resource and another person to provide annotations for feedback and rework.
- standalone tools exist to perform these tasks, they are generally not integrated and are difficult for personnel in different roles to utilize.
- Embodiments of the invention generally are directed at project management related to the production, processing or conversion of motion pictures.
- Large motion picture projects generally utilize workers of several roles to process each image that makes up a motion picture, which may number in the hundreds of thousands of image frames.
- One or more embodiments of the invention enables a computer and database to be configured to accept the assignment of tasks related to artists, time entries for tasks by artists and review of time and actuals of artists by coordinators, a.k.a., "production" and the review of work product by editorial roles.
- the system thus enables artists working on shots made up of multiple images to be managed for successful on-budget completion of projects, along with minimization of generally vast storage requirements for motion picture assets and enable prediction of costs for future bidding on projects given quality, ratings of workers to use and schedule.
- Tasks involved in a motion picture project generally include tasks related to assessment of a project, ingress of a project, assignment of tasks, performing the assigned tasks or project work, reviewing the work product and archiving and shipping the work product of the project.
- One or more embodiments of the invention enable workers of different "roles” to view project tasks in a manner consistent with and which aids their role. This is unknown in the motion picture industry. Roles may be utilized in one or more embodiments of the invention that include "editorial”, “asset manager”, “visual effects supervisor”, “coordinator” or “production”, “artist”, “lead artist”, “stereographer”, “compositor”, “reviewer”, “production assistant”.
- three general categories relate to production workers that manage artists, artist workers who perform the vast majority of work product related work and editorial workers who review and provide feedback based on the work product.
- Each of these roles may utilize a unique or shared view of the motion picture image frames and/or information related to each image or other asset that their role is assigned to work on.
- the editorial and/or asset manager and/or visual effects supervisor roles utilize a tool that shows the motion picture on a display of a computer.
- the tool for example enables the various roles involved in this phase to break a motion picture down into scenes or shots to be worked on.
- One such tool includes "FRAME CYCLER®" commercially available from ADOBE®.
- the asset manager enters the various scene breaks and other resource such as alpha masks, computer generated element layers or any other resources associated with scenes in the motion picture into a database.
- Any type of database may be utilized in one or more embodiments of the invention.
- One such tool that may be utilized to store information related to the motion picture and the assets for project management includes the project management database "TACTICTM", which is commercially available from SOUTHPAW TECHNOLOGY Inc.TM.
- Any database may be utilized in one or more embodiments of the invention so long as the motion picture specific features are included in the project management database.
- One or more embodiments of the invention update the "snapshot" and "file” tables in the project management database.
- the schema of the project management database is briefly described in this section and described in more detail in the detailed description section below.
- production workers utilize an interface that couples with project management database to assign particular workers to particular tasks associated with their role and assign the workers images associated with shots or scenes in a given motion picture.
- One or more embodiments of the invention make use of basic project management database digital asset management tables and add additional fields that improve upon basic project management functionality to optimize the project management process for the motion picture industry.
- One or more embodiments of the invention update the "task" table in the project management database.
- artists, stereographers and compositors perform a large portion of the total work on a motion picture. These roles generally utilize a time clock tool to obtain their tasks and set task status and start and stop times for the task.
- artists perform mask and region design and initial depth augmentation of a frame.
- the artists generally utilize a ray tracing program that may include automated mask tracking capabilities for example, along with NUKETM commercially available from "THE FOUNDRYTM, for mask cleanup for example.
- NUKETM commercially available from "THE FOUNDRYTM
- compositors finish the scene with the same tools that the artists use and generally with other tools such as AFTER EFFECTS® and PHOTOSHOP®, commercially available from ADOBE®.
- the person who worked on a particular asset is stored in the project management database in custom fields for example.
- workers in region design classify elements in scenes into two separate categories.
- Scenes generally include two or more images in time sequence for example.
- the two categories include background elements (i.e. sets and foreground elements that are stationary) or motion elements (e.g., actors, automobiles, etc.) that move throughout the scene.
- background elements i.e. sets and foreground elements that are stationary
- motion elements e.g., actors, automobiles, etc.
- These background elements and motion elements are treated separately in embodiments of the invention similar to the manner in which traditional animation is produced.
- many movies now include computer-generated elements (also known as computer graphics or CG, or also as computer-generated imagery or CGI) that include objects that do not exist in reality, such as robots or spaceships for example, or which are added as effects to movies, for example dust, fog, clouds, etc.
- Computer-generated elements may include background elements, or motion elements.
- Motion Elements The motion elements are displayed as a series of sequential tiled frame sets or thumbnail images complete with background elements.
- the motion elements are masked in a key frame using a multitude of operator interface tools common to paint systems as well as unique tools such as relative bimodal thresholding in which masks are applied selectively to contiguous light or dark areas bifurcated by a cursor brush.
- the mask information from the key frame is then applied to all frames in the display- using mask fitting techniques that include:
- computer-generated elements are imported using RGBAZ files that include an optional alpha mask and/or depths on a pixel-by-pixel, or sub- pixel-by-sub-pixel basis for a computer-generated element.
- this type of file include the EXR file format. Any other file format capable of importing depth and/or alpha information is in keeping with the spirit of the invention.
- Embodiments of the invention import any type of file associated with a computer-generated element to provide instant depth values for a portion of an image associated with a computer-generated element.
- embodiments of the invention allow for the translation, scaling or normalization of the depths for example imported from an RGBAZ file that are associated with computer-generated objects so as to maintain the relative integrity of depth for all of the elements in a frame or sequence of frames.
- any other metadata such as character mattes or alphas or other masks that exist for elements of the images that make up a movie can also be imported and utilized to improve the operated-defined masks used for conversion.
- On format of a file that may be imported to obtain metadata for photographic elements in a scene includes the RGBA file format.
- a final pair of depth enhanced images is thus created based on the input image and any computer-generated element metadata.
- these background elements and motion elements are combined separately into single frame representations of multiple frames, as tiled frame sets or as a single frame composite of all elements (i.e., including both motion and backgrounds/foregrounds) that then becomes a visual reference database for the computer controlled application of masks within a sequence composed of a multiplicity of frames.
- Each pixel address within the reference visual database corresponds to mask/lookup table address within the digital frame and X, Y, Z location of subsequent "raw" frames that were used to create the reference visual database.
- Masks are applied to subsequent frames based on various differentiating image processing methods such as edge detection combined with pattern recognition and other sub-mask analysis, aided by operator segmented regions of interest from reference objects or frames, and operator directed detection of subsequent regions corresponding to the original region of interest.
- image processing methods such as edge detection combined with pattern recognition and other sub-mask analysis, aided by operator segmented regions of interest from reference objects or frames, and operator directed detection of subsequent regions corresponding to the original region of interest.
- the gray scale actively determines the location and shape of each mask (and corresponding color lookup from frame to frame for colorization projects or depth information for two-dimensional to three-dimensional conversion projects) that is applied in a keying fashion within predetermined and operator-controlled regions of interest.
- Camera Pan Background and Static Foreground Elements Stationary foreground and background elements in a plurality of sequential images comprising a camera pan are combined and fitted together using a series of phase correlation, image fitting and focal length estimation techniques to create a composite single frame that represents the series of images used in its construction. During the process of this construction the motion elements are removed through operator adjusted global placement of overlapping sequential frames.
- the single background image representing the series of camera pan images is color designed using multiple color transform look up tables limited only by the number of pixels in the display. This allows the designer to include as much detail as desired including air brushing of mask information and other mask application techniques that provide maximum creative expression.
- the single background image representing the series of camera pan images may be utilized to set depths of the various items in the background. Once the background color/depth design is completed the mask information is transferred automatically to all the frames that were used to create the single composited image. In this manner, color or depth is performed once per multiple images and/or scene instead of once per frame, with color/depth information automatically spread to individual frames via embodiments of the invention.
- Masks from colorization projects may be combined or grouped for depth conversion projects since the colorization masks may contain more sub-areas than a depth conversion mask. For example, for a coloration project, a person's face may have several masks applied to areas such as lips, eyes, hair, while a depth conversion project may only require an outline of the person's head or an outline of a person's nose, or a few geometric shape sub-masks to which to apply depth.
- Masks from a colorization project can be utilized as a starting point for a depth conversion project since defining the outlines of human recognizable objects by itself is time consuming and can be utilized to start the depth conversion masking process to save time. Any computer-generated elements at the background level may be applied to the single background image.
- image offset information relative to each frame is registered in a text file during the creation of the single composite image representing the pan and used to apply the single composite mask to frames used to create the composite image.
- the instance of background and foreground previously occluded is copied into the single image with priority and proper offsets to compensate for camera movement.
- the offset information is included in a text file associated with each single representation of the background so that the resulting mask information can be applied to each frame in the scene with proper mask offsets.
- the single background image representing the series of static camera frames is color designed using multiple color transform look up tables limited only by the number of pixels in the display. Where the motion elements occlude the background elements continuously within the series of sequential frames they are seen as black figure that are ignored and masked over. The black objects are ignored in colorization-only projects during the masking operation because the resulting background mask is later applied to all frames used to create the single representation of the background only where there is no pre-existing mask. If background information is created for areas that are never exposed, then this data is treated as any other background data that is spread through a series of images based on the composite background.
- the distance from the camera to each item in the composite frame is automatically transferred to all the frames that were used to create the single composited image.
- This horizontal shifting may utilize data generated by an artist for the occluded or alternatively, areas where no image data exists yet for a second viewpoint may be marked in one or more embodiments of the invention using a user defined color that allows for the creation missing data to ensure that no artifacts occur during the two-dimension to three-dimension conversion process.
- Any technique known may be utilized in embodiments of the invention to cover areas in the background where unknown data exists, i.e., (as displayed in some color that shows where the missing data exists) that may not be borrowed from another scene/frame for example by having artists create complete backgrounds or smaller occluded areas with artist drawn objects.
- a second viewpoint image may be created for each image in a scene in order to produce a stereoscopic view of the movie, for example a left eye view where the original frames in the scene are assigned to the right eye viewpoint, for example by translating foreground objects horizontally for the second viewpoint, or alternatively by translating foreground objects horizontally left and right to create two viewpoints offset from the original viewpoint.
- One or more tools employed by the system enable real-time editing of 3D images without re-rendering for example to alter layers/colors/masks and/or remove artifacts and to minimize or eliminate iterative workflow paths back through different workgroups by generating translation files that can be utilized as portable pixel-wise editing files.
- a mask group takes source images and creates masks for items, areas or human recognizable objects in each frame of a sequence of images that make up a movie.
- the depth augmentation group applies depths, and for example shapes, to the masks created by the mask group.
- left and right viewpoint images and left and right translation files may be generated by one or more embodiments of the invention.
- the left and right viewpoint images allow 3D viewing of the original 2D image.
- the translation files specify the pixel offsets for each source pixel in the original 2D image, for example in the form of UV or U maps. These files are generally related to an alpha mask for each layer, for example a layer for an actress, a layer for a door, a layer for a background, etc. These translation files, or maps are passed from the depth augmentation group that renders 3D images, to the quality assurance workgroup.
- the Z depth of regions within the image may also be passed along with the alpha mask to the quality assurance group, who may then adjust depth as well without re-rendering with the original rendering software. This may be performed for example with generated missing background data from any layer so as to allow "downstream" real-time editing without re-rendering or ray-tracing for example.
- Quality assurance may give feedback to the masking group or depth augmentation group for individuals so that these individuals may be instructed to produce work product as desired for the given project, without waiting for, or requiring the upstream groups to rework anything for the current project.
- the asset is reviewed for example using an interface that couples with the project management database to enable the viewing of work product.
- editorial role based users use the interface most, artists and stereographers less and lead artists the least.
- the review notes and images may be viewed simultaneously, for example with a clear background surrounding text that is overlaid on the image or scene to enable rapid review and feedback by a given worker having a particular role.
- Other improvements to the project management database include ratings or artists and difficulty of the asset.
- Asset managers may delete and/or compress all assets that may be regenerated, which can save hundreds of terabytes of disk space for a typical motion picture. This enables an enormous savings in disk drive hardware purchases and is unknown in the art.
- One or more embodiments of the system may be implemented with a computer and a database coupled with the computer.
- Any computer architecture having any number of computers, for example coupled via a computer communication network is in keeping with the spirit of the invention.
- the database coupled with the computer includes at least a project table, shot table, task table and timesheet table.
- the project table generally includes project identifier and description of a project related to a motion picture.
- the shot table generally includes a shot identifier and references a plurality of images with a starting frame value and an ending frame value wherein the plurality of images are associated with the motion picture that is associated with the project.
- the shot table generally includes at least one shot having status related to progress of work performed on the shot.
- the task table generally references the project using a project identifier in also located in the project table.
- the task table generally includes at least one task which generally includes a task identifier and an assigned worker, e.g., artist, and which may also include a context setting associated with a type of task related to motion picture work selected from region design, setup, motion, composite, and review.
- the at least one task generally includes a time allocated to complete the at least one task.
- the timesheet item table generally references the project identifier in the project table and the task identifier in the task table.
- the task table generally includes at least one timesheet item that includes a start time and an end time.
- the computer is configured to present a first display configured to be viewed by an artist that includes at least one daily assignment having a context, project, shot and a status input that is configured to update the status in the task table and a timer input that is configured to update the start time and the end time in the timesheet item table.
- the computer is generally configured to present a second display configured to be viewed by a coordinator or "production" worker, i.e., production that includes a search display having a context, project, shot, status and artist and wherein the second display further includes a list of a plurality of artists and respective status and actuals based on time spent in the at least one timesheet item versus the time allocated per the at least one task associated with the at least one shot.
- the computer is generally also configured to present a third display configured to be viewed by an editor that includes an annotation frame configured to accept commentary or drawing or both commentary and drawing on the at least one of said plurality of images associated with the at least one shot.
- One or more embodiments of the computer may be configured to provide the third display configured to be viewed by an editor that includes an annotation overlaid on at least one of the plurality of images. This capability provides information on one display that has generally required three workers to integrate in known systems, and which is novel in and of itself.
- Embodiments of the database may also include a snapshot table which includes an snapshot identifier and search type and which includes a snapshot of the at least one shot, for example that includes a subset of the at least one shot wherein the snapshot is cached on the computer to reduce access to the shot table.
- Embodiments may also include other context settings for other types of task categories, for example source and cleanup related tasks. Any other context settings or values that are related to motion picture work may also be included in keeping with the spirit of the invention.
- Embodiments of the database may also include an asset request table that includes an asset request identifier and shot identifier that may be utilized to request work on assets or assets themselves to be worked on or created by other workers for example.
- Embodiments of the database may also include a request table that includes an mask request identifier and shot identifier and that may be utilized to request any type of action by another worker for example.
- Embodiments of the database may also include a note table which includes a note identifier and that references the project identifier and that includes at least one note related to at least one of the plurality of images from the motion picture.
- Embodiments of the database may also include a delivery table that includes a delivery identifier that references the project identifier and which includes information related to delivery of the motion picture.
- One or more embodiments of the computer are configured to accept a rating input from production or the editor based on work performed by the artist, optionally in a blind manner in which the reviewer does not know the identity of the artist in order to prevent favoritism for example.
- One or more embodiments of the computer are configured to accept a difficulty of the at least one shot and calculate a rating based on work performed by the artist and based on the difficulty of the shot and time spent on the shot.
- One or more embodiments of the computer are configured to accept a rating input from production or editorial, (i.e., an editor worker) based on work performed by the artist, or, accept a difficulty of the at least one shot and calculate a rating based on work performed by the artist and based on the difficulty of the shot and time spent on the shot, and, signify an incentive with respect to the artist based on the rating accepted by the computer or calculated by the computer.
- One or more embodiments of the computer are configured to estimate remaining cost based on the actuals that are based on total time spent for all of the at least one tasks associated with all of the at least one shot in the project versus time allocated for all of the at least one tasks associated with all of the at least one shot in the project.
- One or more embodiments of the computer are configured to compare the actuals associated with a first project with actuals associated with a second project and signify at least one worker to be assigned from the first project to the second project based on at least one rating of the first worker that is assigned to the first project.
- One or more embodiments of the computer are configured to analyze a prospective project having a number of shots and estimated difficulty per shot and based on actuals associated with the project, calculate a predicted cost for the prospective project.
- One or more embodiments of the computer are configured to analyze a prospective project having a number of shots and estimated difficulty per shot and based on the actuals associated with a first previously performed project and a second previously performed project that completed after the first previously performed project, calculate a derivate of the actuals, calculate a predicted cost for the prospective project based on the derivative of the actuals. For example, as the process improves, tools improve and workers improve, the efficiency of work improves and the budgeting and bid processes can take this into account by calculating how efficiency is changing versus time and use this rate of change to predict costs for a prospective project.
- One or more embodiments of the computer are configured to analyze the actuals associated with said project and divide completed shots by total shots associated with said project and present a time of completion of the project.
- One or more embodiments of the computer are configured to analyze the actuals associated with the project and divide completed shots by total shots associated with the project, present a time of completion of the project, accept an input of at least one additional artist having a rating, accept a number of shots in which to use the additional artist, calculate a time savings based on the at least one additional artist and the number of shots, subtract the time savings from the time of completion of the project and present an updated time of completion of the project.
- One or more embodiments of the computer are configured to calculate amount of disk space that may be utilized to archive the project and signify at least one asset that may be rebuilt from other assets to avoid archival of the at least one asset.
- One or more embodiments of the computer are configured to display an error message if the artist is working with a frame number that is not current in the at least one shot. This may occur when fades, dissolves or other effects lengthen a particular shot for example wherein the shot contains frames not in the original source assets.
- FIG. 1 shows a plurality of feature film or television film frames representing a scene or cut in which there is a single instance or perceptive of a background.
- FIG. 2 shows an isolated background processed scene from the plurality of frames shown in FIG. 1 in which all motion elements are removed using various subtraction and differencing techniques.
- the single background image is then used to create a background mask overlay representing designer selected color lookup tables in which dynamic pixel colors automatically compensate or adjust for moving shadows and other changes in luminance.
- FIG. 3 shows a representative sample of each motion object (M-Object) in the scene receives a mask overlay that represents designer selected color lookup tables in which dynamic pixel colors automatically compensate or adjust for moving shadows and other changes in luminance as the M-Object moves within the scene.
- FIG. 4 shows all mask elements of the scene are then rendered to create a fully colored frame in which M-Object masks are applied to each appropriate frame in the scene followed by the background mask, which is applied only where there is no pre-existing mask in a Boolean manner.
- FIGS. 5 A and 5B show a series of sequential frames loaded into display memory in which one frame is fully masked with the background (key frame) and ready for mask propagation to the subsequent frames via automatic mask fitting methods.
- FIGS. 6 A and 6B show the child window displaying an enlarged and scalable single image of the series of sequential images in display memory.
- the Child window enables the operator to manipulate masks interactively on a single frame or in multiple frames during real time or slowed motion.
- FIGS. 7 A and 7B shows a single mask (flesh) is propagated automatically to all frames in the display memory.
- FIG. 8 shows all masks associated with the motion object are propagated to all sequential frames in display memory.
- FIG. 9A shows a picture of a face.
- FIG. 9B shows a close up of the face in FIG. 9A wherein the "small dark" pixels shown in FIG. 9B are used to calculate a weighed index using bilinear interpolation.
- FIGS 10A-D show searching for a Best Fit on the Error Surface:
- An error surface calculation in the Gradient Descent Search method involves calculating mean squared differences of pixels in the square fit box centered on reference image pixel (xO, yO), between the reference image frame and the corresponding (offset) location (x, y) on the search image frame.
- FIGS. 11 A-C show a second search box derived from a descent down the error surface gradient (evaluated separately), for which the evaluated error function is reduced, possibly minimized, with respect to the original reference box (evident from visual comparison of the boxes with the reference box in FIGS. 10A, B, C and D.).
- FIG. 12 depicts the gradient component evaluation.
- the error surface gradient is calculated as per definition of the gradient.
- Vertical and horizontal error deviations are evaluated at four positions near the search box center position, and combined to provide an estimate of the error gradient for that position. 12.
- FIG. 13 shows a propagated mask in the first sequential instance where there is little discrepancy between the underlying image data and the mask data.
- the dress mask and hand mask can be clearly seen to be off relative to the image data.
- FIG. 14 shows that by using the automatic mask fitting routine, the mask data adjusts to the image data by referencing the underlying image data in the preceding image.
- FIG. 15 shows the mask data in later images within the sequence show marked discrepancy relative to the underlying image data. Eye makeup, lipstick, blush, hair, face, dress and hand image data are all displaced relative to the mask data.
- FIG. 16 shows that the mask data is adjusted automatically based on the underlying image data from the previous mask and underlying image data.
- FIG. 17 shows the mask data from FIG. 16 is shown with appropriate color transforms after whole frame automatic mask fitting. The mask data is adjusted to fit the underlying luminance pattern based on data from the previous frame or from the initial key frame.
- FIG. 18 shows polygons that are used to outline a region of interest for masking in frame one.
- the square polygon points snap to the edges of the object of interest.
- the Bezier points snap to the object of interest and the control points/curves shape to the edges.
- FIG. 19 shows the entire polygon or Bezier curve is carried to a selected last frame in the display memory where the operator adjusts the polygon points or Bezier points and curves using the snap function which automatically snaps the points and curves to the edges of the object of interest.
- FIG. 20 shows that if there is a marked discrepancy between the points and curves in frames between the two frames where there was an operator interactive adjustment, the operator will further adjust a frame in the middle of the plurality of frames where there is maximum error of fit.
- FIG. 21 shows that when it is determined that the polygons or Bezier curves are correctly animating between the two adjusted frames, the appropriate masks are applied to all frames.
- FIG. 22 shows the resulting masks from a polygon or Bezier animation with automatic point and curve snap to edges.
- the brown masks are the color transforms and the green masks are the arbitrary color masks.
- FIG 23 shows an example of two pass blending:
- the objective in two-pass blending is to eliminate moving objects from the final blended mosaic. This can be done by first blending the frames so the moving object is completely removed from the left side of the background mosaic. As shown in FIG. 23, the character can is removed from the scene, but can still be seen in the right side of the background mosaic.
- FIG. 24 shows the second pass blend.
- a second background mosaic is then generated, where the blend position and width is used so that the moving object is removed from the right side of the final background mosaic.
- the character can is removed from the scene, but can still be seen the left side of the background mosaic.
- the moving character is shown on the left.
- FIG. 25 shows the final background corresponding to FIGS. 23-24.
- the two-passes are blended together to generate the final blended background mosaic with the moving object removed from the scene.
- the final blended background with moving character is removed.
- FIG. 26 shows an edit frame pair window.
- FIG. 27 shows sequential frames representing a camera pan that are loaded into memory.
- the motion object (butler moving left to the door) has been masked with a series of color transform information leaving the background black and white with no masks or color transform information applied.
- FIG. 28 shows six representative sequential frames of the pan above are displayed for clarity.
- FIG. 29 shows the composite or montage image of the entire camera pan that was built using phase correlation techniques.
- the motion object (butler) included as a transparency for reference by keeping the first and last frame and averaging the phase correlation in two directions.
- the single montage representation of the pan is color designed using the same color transform masking techniques as used for the foreground object.
- FIG. 30 shows that the sequence of frames in the camera pan after the background mask color transforms the montage has been applied to each frame used to create the montage.
- the mask is applied where there is no pre-existing mask thus retaining the motion object mask and color transform information while applying the background information with appropriate offsets.
- FIG. 31 shows a selected sequence of frames in the pan for clarity after the color background masks have been automatically applied to the frames where there is no pre-existing masks.
- FIG. 32 shows a sequence of frames in which all moving objects (actors) are masked with separate color transforms.
- FIG. 33 shows a sequence of selected frames for clarity prior to background mask information. All motion elements have been fully masked using the automatic mask-fitting algorithm.
- FIG. 34 shows the stationary background and foreground information minus the previously masked moving objects.
- the single representation of the complete background has been masked with color transforms in a manner similar to the motion objects.
- outlines of removed foreground objects appear truncated and unrecognizable due to their motion across the input frame sequence interval, i.e., the black objects in the frame represent areas in which the motion objects (actors) never expose the background and foreground.
- the black objects are ignored during the masking operation in colorization-only projects because the resulting background mask is later applied to all frames used to create the single representation of the background only where there is no pre-existing mask.
- the missing data area may be displayed so that image data may be obtained/generated for the missing data area so as to provide visually believable image data when translating foreground objects horizontally to generate a second viewpoint.
- FIG. 35 shows the sequential frames in the static camera scene cut after the background mask information has been applied to each frame with appropriate offsets and where there is no pre-existing mask information.
- FIG. 36 shows a representative sample of frames from the static camera scene cut after the background information has been applied with appropriate offsets and where there is no preexisting mask information.
- FIGS. 37A-C show embodiments of the Mask Fitting functions, including calculate fit grid and interpolate mask on fit grid.
- FIGS. 38A-B show embodiments of the extract background functions.
- FIGS. 39A-C show embodiments of the snap point functions.
- FIGS. 40A-C show embodiments of the bimodal threshold masking functions, wherein FIG. 40C corresponds to step 2.1 in FIG. 40A, namely "Create Image of Light/Dark Cursor Shape" and FIG. 40B corresponds to step 2.2 in Figure 40A, namely "Apply Light/Dark shape to
- FIGS. 41A-B show embodiments of the calculate fit value functions.
- FIG. 42 shows two image frames that are separated in time by several frames, of a person levitating a crystal ball wherein the various objects in the image frames are to be converted from two-dimensional objects to three-dimensional objects.
- FIG. 43 shows the masking of the first object in the first image frame that is to be converted from a two-dimensional image to a three-dimensional image.
- FIG. 44 shows the masking of the second object in the first image frame.
- FIG. 45 shows the two masks in color in the first image frame allowing for the portions associated with the masks to be viewed.
- FIG. 46 shows the masking of the third object in the first image frame.
- FIG. 47 shows the three masks in color in the first image frame allowing for the portions associated with the masks to be viewed.
- FIG. 48 shows the masking of the fourth object in the first image frame.
- FIG. 49 shows the masking of the fifth object in the first image frame.
- FIG. 50 shows a control panel for the creation of three-dimensional images, including the association of layers and three-dimensional objects to masks within an image frame, specifically showing the creation of a Plane layer for the sleeve of the person in the image.
- FIG. 51 shows a three-dimensional view of the various masks shown in FIGS. 43-49, wherein the mask associated with the sleeve of the person is shown as a Plane layer that is rotated toward the left and right viewpoints on the right of the page.
- FIG. 52 shows a slightly rotated view of FIG. 51.
- FIG. 53 shows a slightly rotated view of FIG. 51.
- FIG. 54 shows a control panel specifically showing the creation of a sphere object for the crystal ball in front of the person in the image.
- FIG. 55 shows the application of the sphere object to the flat mask of the crystal ball, that is shown within the sphere and as projected to the front and back of the sphere to show the depth assigned to the crystal ball.
- FIG. 56 shows a top view of the three-dimensional representation of the first image frame showing the Z-dimension assigned to the crystal ball shows that the crystal ball is in front of the person in the scene.
- FIG. 57 shows that the sleeve plane rotating in the X-axis to make the sleeve appear to be coming out of the image more.
- FIG. 58 shows a control panel specifically showing the creation of a Head object for application to the person's face in the image, i.e., to give the person's face realistic depth without requiring a wire model for example.
- FIG. 59 shows the Head object in the three-dimensional view, too large and not aligned with the actual person's head.
- FIG. 60 shows the Head object in the three-dimensional view, resized to fit the person's face and aligned, e.g., translated to the position of the actual person's head.
- FIG. 61 shows the Head object in the three-dimensional view, with the Y-axis rotation shown by the circle and Y-axis originating from the person's head thus allowing for the correct rotation of the Head object to correspond to the orientation of the person's face.
- FIG. 62 shows the Head object also rotated slightly clockwise, about the Z-axis to correspond to the person's slightly tilted head.
- FIG. 63 shows the propagation of the masks into the second and final image frame.
- FIG. 64 shows the original position of the mask corresponding to the person's hand.
- FIG. 65 shows the reshaping of the mask, that can be performed automatically and/or manually, wherein any intermediate frames get the tweened depth information between the first image frame masks and the second image frame masks.
- FIG. 66 shows the missing information for the left viewpoint as highlighted in color on the left side of the masked objects in the lower image when the foreground object, here a crystal ball is translated to the right.
- FIG. 67 shows the missing information for the right viewpoint as highlighted in color on the right side of the masked objects in the lower image when the foreground object, here a crystal ball is translated to the left.
- FIG. 68 shows an anaglyph of the final depth enhanced first image frame viewable with Red/Blue 3-D glasses.
- FIG. 69 shows an anaglyph of the final depth enhanced second and last image frame viewable with Red/Blue 3-D glasses, note rotation of person's head, movement of person's hand and movement of crystal ball.
- FIG. 70 shows the right side of the crystal ball with fill mode "smear", wherein the pixels with missing information for the left viewpoint, i.e., on the right side of the crystal ball are taken from the right edge of the missing image pixels and "smeared" horizontally to cover the missing information.
- FIG. 71 shows a mask or alpha plane, for an actor's upper torso and head (and transparent wings).
- the mask may include opaque areas shown as black and transparent areas that are shown as grey areas.
- FIG. 72 shows an occluded area, that corresponds to the actor of FIG. 71, and that shows an area of the background that is never exposed in any frame in a scene. This may be a composite background for example.
- FIG. 73 shows the occluded area artistically rendered to generate a complete and realistic background for use in two-dimensional to three-dimensional conversion, so as to enable an artifact-free conversion.
- FIG. 73 A shows the occluded area partially drawn or otherwise rendered to generate just enough of a realistic looking background for use in minimizing artifacts two-dimensional to three-dimensional conversion.
- FIG. 74 shows a light area of the shoulder portion on the right side of FIG. 71 that represents a gap where stretching (as is also shown in FIG. 70) would be used when shifting the foreground object to the left to create a right viewpoint.
- the dark portion of the figure is taken from the background where data is available in at least one frame of a scene.
- FIG. 75 shows an example of the stretching of pixels, i.e., smearing, corresponding to the light area in FIG. 74 without the use of a generated background, i.e., if no background data is available for an area that is occluded in all frames of a scene.
- FIG. 76 shows a result of a right viewpoint without artifacts on the edge of the shoulder of the person wherein the dark area includes pixels available in one or more frames of a scene, and generated data for always-occluded areas of a scene.
- FIG. 77 shows an example of a computer-generated element, here a robot, which is modeled in three-dimensional space and projected as a two-dimensional image. If metadata such as alpha, mask, depth or any combination thereof exists, the metadata can be utilized to speed the conversion process from two-dimensional image to a pair of two-dimensional images for left and right eye for three-dimensional viewing.
- metadata such as alpha, mask, depth or any combination thereof exists
- FIG. 78 shows an original image separated into a background and foreground elements, (mountain and sky in the background and soldiers in the bottom left also see Fig. 79) along with the imported color and depth of the computer-generated element, i.e., the robot with depth automatically set via the imported depth metadata.
- any area that is covered for the scene can be artistically rendered for example to provide believable missing data, as is shown in Fig. 73 based on the missing data of Fig. 73 A, which results in artifact free edges as shown in Fig. 76 for example.
- FIG. 79 shows masks associated with the photograph of soldiers in the foreground to apply depth to the various portions of the soldiers that lie in depth in front of the computer- generated element, i.e., the robot.
- the dashed lines horizontally extending from the mask areas show horizontal translation of the foreground objects takes place and where imported metadata can be utilized to accurately auto-correct over-painting of depth or color on the masked objects when metadata exists for the other elements of a movie. For example, when an alpha exists for the objects that occur in front of the computer-generated elements.
- One type of file that can be utilized to obtain mask edge data is a file with alpha file and/or mask data such as an RGBA file.
- FIG. 80 shows an imported alpha layer which can also be utilized as a mask layer to limit the operator defined, and potentially less accurate masks used for applying depth to the edges of the three soldiers A, B and C.
- a computer-generated element for dust can be inserted into the scene along the line annotated as "DUST", to augment the reality of the scene.
- FIG. 81 shows the result of using the operator-defined masks without adjustment when overlaying a motion element such as the soldier on the computer-generated element such as the robot.
- a motion element such as the soldier on the computer-generated element such as the robot.
- FIG. 82 shows a source image to be depth enhanced and provided along with left and right translation files and alpha masks so that downstream workgroups may perform real-time editing of 3D images without re-rendering for example to alter layers/colors/masks and/or remove and/or or adjust depths without iterative workflow paths back to the original workgroups.
- FIG. 83 shows masks generated by the mask workgroup for the application of depth by the depth augmentation group, wherein the masks are associated with objects, such as for example human recognizable objects in the source image of FIG. 82.
- FIG. 84 shows areas where depth is applied generally as darker for nearer objects and lighter for objects that are further away.
- FIG. 85A shows a left UV map containing translations or offsets in the horizontal direction for each source pixel.
- FIG. 85B shows a right UV map containing translations or offsets in the horizontal direction for each source pixel.
- FIG. 85C shows a black value shifted portion of the left UV map of FIG. 85A to show the subtle contents therein.
- FIG. 85D shows a black value shifted portion of the right UV map of FIG. 85B to show the subtle contents therein.
- FIG. 86A shows a left U map containing translations or offsets in the horizontal direction for each source pixel.
- FIG. 86B shows a right U map containing translations or offsets in the horizontal direction for each source pixel.
- FIG. 86C shows a black value shifted portion of the left U map of FIG. 86A to show the subtle contents therein.
- FIG. 86D shows a black value shifted portion of the right U map of FIG. 86B to show the subtle contents therein.
- FIG. 87 shows known uses for UV maps, wherein a three-dimensional model is unfolded so that an image in UV space can be painted onto the 3D model using the UV map.
- FIG. 88 shows a disparity map showing the areas where the difference between the left and right translation maps is the largest.
- FIG. 89 shows a left eye rendering of the source image of FIG. 82.
- FIG. 90 shows a right eye rendering of the source image of FIG. 82.
- FIG. 91 shows an anaglyph of the images of FIG. 89 and FIG. 90 for use with Red/Blue glasses.
- FIG. 92 shows an image that has been masked and is in the process of depth enhancement for the various layers.
- FIG. 93 shows a UV map overlaid onto an alpha mask associated with the actress shown in FIG. 92 which sets the translation offsets in the resulting left and right UV maps based on the depth settings of the various pixels in the alpha mask.
- FIG. 94 shows a workspace generated for a second depth enhancement program, or compositing program such as NUKE ®, i.e., generated for the various layers shown in FIG. 92, i.e., left and right UV translation maps for each of the alphas wherein the workspace allows for quality assurance personnel (or other work groups) to perform real-time editing of 3D images without re-rendering for example to alter layers/colors/masks and/or remove artifacts or otherwise adjust masks and hence alter the 3D image pair (or anaglyph) without iteratively sending fixes to any other workgroup.
- a second depth enhancement program such as NUKE ®
- FIG. 95 shows a workflow for iterative corrective workflow.
- FIG. 96 shows an embodiment of the workflow enabled by one or more embodiments of the system in that each workgroup can perform real-time editing of 3D images without re- rendering for example to alter layers/colors/masks and/or remove artifacts and otherwise correct work product from another workgroup without iterative delays associated with re -rendering/ray- tracing or sending work product back through the workflow for corrections.
- FIG. 97 illustrates and architectural view of an embodiment of the invention.
- FIG. 98 illustrates an annotated view of a session manager window utilized to define a plurality of images to work on or assign work to.
- FIG. 99 illustrates a view of the production display showing a project, shots and tasks related to the selected shot, along with status for each task context associated with the shot.
- FIG. 100 illustrates a view of the actuals associated with a particular shot within a project for each task context associated with the shot wherein "under bid" task actuals are shown in a first manner and tasks that are within a predefined percentage within the bid amount are shown in a second manner while tasks that are over bid are shown in a third manner.
- FIG. 101 illustrates the amount of disk space that may be saved by deleting files which can be reconstructed from other files, for example after completion of a project to save disk drive expenditures.
- FIG. 102 illustrates a view of an artist display showing the task context, project, shot, status, tools, start time button, check-in input, rendering input, internal shot review input, meal, start/time/stop, review and submit inputs.
- FIG. 103 illustrates an annotated view of the menu bar of the artist display.
- FIG. 104 illustrates an annotated view of the task row of the artist display.
- FIG. 105 illustrates an annotated view of the main portion of the user interface of the artist display.
- FIG. 106 illustrates a build timeline display for the artist display for creating a timeline to work on.
- FIG. 107 illustrates a browse snapshots display for the artist display that enables an artist to view the snapshot of a shot or otherwise cache important information related to a shot so that the database does not have to field requests for often utilized data.
- FIG. 108 illustrates an artist actual window showing the actuals of time spent on tasks, for example versus time allocated for the tasks, with dropdown menus for specific timesheets.
- FIG. 109 illustrates a notes display for the artist display that enables artists to enter a note related to the shot.
- FIG. 110 illustrates a check in display for the artist display that enables work to be checked in after work on the shot is complete.
- FIG. I l l illustrates a view of an editorial display showing the project, filter inputs, timeline inputs and search results showing the shots in the main portion of the window along with the context of work and assigned worker.
- FIG. 112 illustrates a view of the session manager display of the editorial display for selecting shots to review.
- FIG. 113 illustrates a view of the advanced search display of the editorial display.
- FIG. 114 illustrates a view of the simple search display of the editorial display.
- FIG. 115 illustrates a view of the reviewing pane for a shot also showing integrated notes and/or snapshot information in the same frame.
- FIG. 116 illustrates a view of the timelines to select for review and/or checkin after modification.
- FIG. 117 illustrates the annotations added to a frame for feedback using the tool of FIG. 117.
- FIG. 97 illustrates and architectural view of an embodiment of the invention.
- One or more embodiments of the system include computer 9702 and database 9701 coupled with computer 9702. Any computer architecture having any number of computers, for example coupled via a computer communication network is in keeping with the spirit of the invention.
- Database 9701 coupled with computer 9702 includes at least a project table, shot table, task table and timesheet table.
- the project table generally includes project identifier and description of a project related to a motion picture.
- the shot table generally includes a shot identifier and references a plurality of images with a starting frame value and an ending frame value wherein the plurality of images are associated with the motion picture that is associated with the project.
- the shot table generally includes at least one shot having status related to progress of work performed on the shot.
- the task table generally references the project using a project identifier in also located in the project table.
- the task table generally includes at least one task which generally includes a task identifier and an assigned worker, e.g., artist, and which may also include a context setting associated with a type of task related to motion picture work selected from region design, setup, motion, composite, and review for example (or any other set of motion picture related types of tasks).
- the context setting may also imply or have a default workflow so that region design flows into depth which flows into composite. This enables the system to assign the next type of task, or context that the shot is to have work performed on. The flow may be linear or may iterate back for rework for example.
- the at least one task generally includes a time allocated to complete the at least one task.
- the timesheet item table generally references the project identifier in the project table and the task identifier in the task table.
- the task table generally includes at least one timesheet item that includes a start time and an end time.
- the completion of tasks may in one or more embodiments set the context of the task to the next task in sequence in a workflow and the system may automatically notify the next worker in the workflow based on the next context of work to be performed and the workers may work under different contexts as previously described.
- contexts may have sub- contexts, i.e., region design may be broken into masking and outsource masking while depth may be broken into key frame and motion contexts, depending on the desired workflow for the specific type of operations to be performed on a motion picture project.
- Embodiments of the database may also include a snapshot table which includes a snapshot identifier and search type and which includes a snapshot of the at least one shot, for example that includes a subset of the at least one shot wherein the snapshot is cached on the computer to reduce access to the shot table.
- Other embodiments of the snapshot table keep track of the resources on the network, stores information about the resources and track versioning of the resource.
- Embodiments may also include other context settings for other types of task categories, for example source and cleanup related tasks. Any other context settings or values that are related to motion picture work may also be included in keeping with the spirit of the invention.
- Embodiments of the database may also include an asset request table that includes an asset request identifier and shot identifier that may be utilized to request work on assets or assets themselves to be worked on or created by other workers for example.
- Embodiments of the database may also include a request table that includes an mask request identifier and shot identifier and that may be utilized to request any type of action by another worker for example.
- Embodiments of the database may also include a note table which includes a note identifier and that references the project identifier and that includes at least one note related to at least one of the plurality of images from the motion picture.
- Embodiments of the database may also include a delivery table that includes a delivery identifier that references the project identifier and which includes information related to delivery of the motion picture.
- One or more embodiments of the database may utilize a schema as follows or any other schema that is capable of supporting the functionality of the invention as specifically programmed in computer 9702 and as described in any combination or sub-combination as follows so long as motion picture project management may be performed as detailed herein or to in any other way better manage motion picture projects using the exemplary specifications herein:
- unique task identifier assigned worker, description (what to do), status (ingested, waiting, complete, returned, approved), bid start date, bid end date, bid duration, actual start date, actual end date, priority, context (stacking, asset, motion, motion visual effects, outsource, cleanup, alpha generation, composite, masking, clean plate, setup, keyframe, quality control), project code in project table, supervisor or production or editorial worker, time spent per process.
- unique note identifier unique note identifier, project code, search type, search id, login (worker id), context (composite, review, motion, editorial%), timestamp, note (text description of note associated with set of images defined by search).
- unique deliver identifier login (of worker), timestamp, status (retired or not), delivery method (how it was delivered), description (what type of media used to deliver project), returned (true or false), drive (serial number of the drive), case (serial number on the case), delivery date, project identifier, client (text name of client), producer (name of producer).
- time sheet unique identifier login (worker), timestamp, total time, timesheet approver, start time, end time, meal 1 (half hour break start time), meal 2 (half hour break start time), status (pending or approved).
- timesheet item unique identifier login (worker), timestamp, context (region design, composite, rendering, motion, management, mask cleanup, training, cleanup, admin), project identifier, timesheet identifier, start time, end time, status (pending or approved), approved by (worker), task identifier.
- sequence unique identifier login (worker that defined sequence), timestamp, shot order (that makes up the sequence).
- shot unique identifier login (worker that defined shot), timestamp, shot status (in progress, final, final client approval), client status (composite in progress, depth client review, composite client review, final), description (text description of shot, e.g., 2 planes flying by each other), first frame number, last frame number, number of frames, assigned worker, region design, depth completion date, depth worker assigned, composite supervisor, composite lead worker, composite completion date.
- asset unique identifier [00191] asset unique identifier, timestamp, asset worker assigned, status (pending or resolved), shot identifier, problem description in text, production worker, lead worker assigned, priority, due date.
- mask request unique identifier login (worker making mask request), timestamp, depth artist, depth lead, depth coordinator or production worker, masking issues, masks (versions with issue related to mask request), source used, due date, rework notes.
- the computer is generally configured to present a session manager to select a series of images within a shot to work on and/or assign tasks to or to review.
- the computer is generally configured to present a first display configured to be viewed by production that includes a search display having a context, project, shot, status and artist and wherein the second display further includes a list of a plurality of artists and respective status and actuals based on time spent in the at least one timesheet item versus the time allocated per the at least one task associated with the at least one shot.
- FIG. 98 illustrates an annotated view of a session manager window utilized to define a plurality of images to work on or assign work to or to review for example.
- Computer 9702 accepts inputs for the project, sequence (the motion picture or a trailer for example that utilizes shots in a particular sequence), along with the shot, mask version and various frame offsets and optionally downloads the images to the local computer for local processing for example.
- sequence the motion picture or a trailer for example that utilizes shots in a particular sequence
- the shot the motion picture or a trailer for example that utilizes shots in a particular sequence
- mask version the motion picture or a trailer for example that utilizes shots in a particular sequence
- various frame offsets optionally downloads the images to the local computer for local processing for example.
- Each field on the figure is annotated with further details.
- FIG. 99 illustrates a view of the production display showing a project, shots and tasks related to the selected shot, along with status for each task context associated with the shot.
- the shots that make up a project may be selected, which redirects the main portion of the window to display information related to the selected shot(s) including tabs for "shot info", "assets” used in the shot, frames to select to view, notes, task information, actuals, check-in and data integrity.
- shots As shown in the main portion of the window, several task contexts are shown with their associated status, assigned workers, etc. Production utilizes this display and the computer accepts inputs from the production worker, e.g., user via this display to set tasks for artists along with allotted times.
- a small view of the shot is shown in the lower left of the display to give production workers a view of the motion picture shot related to the tasks and context settings.
- One potential goal of the production role is to assign tasks and review status and actuals, while one potential goal of the artist is to use a set of tools to manipulate images, while one potential goal of the reviewer is to have a higher resolution image for review with integrated meta data related to the shot and status thereof.
- the displays in the system are tailored for roles, yet integrated to prioritize the relevant information mainly of importance to that role in the motion picture creation/conversion process.
- FIG. 100 illustrates a view of the actuals associated with a particular shot within a project for each task context associated with the shot wherein "under bid" task actuals are shown in a first manner and tasks that are within a predefined percentage within the bid amount are shown in a second manner while tasks that are over bid are shown in a third manner.
- FIG. 101 illustrates the amount of disk space that may be saved by deleting files which can be reconstructed from other files, for example after completion of a project to save disk drive expenditures.
- rebuildable amounts for a project only partially complete can be in the terabyte range easily.
- the computer accesses the database and determines which resources depend on other resources and whether they can be compressed and with what general compression ratio that may be calculated in advance and/or based on other projects for example.
- the computer calculates the total storage and the amount of storage that may be freed up by compression and/or regeneration of resources and displays the information on a computer display for example.
- the computer is also generally configured to present a second display configured to be viewed by an artist that includes at least one daily assignment having a context, project, shot and a status input that is configured to update the status in the task table and a timer input that is configured to update the start time and the end time in the timesheet item table.
- FIG. 102 illustrates a view of an artist display showing the task context, project, shot, status, tools, start time button, check-in input, rendering input, internal shot review input, meal, start/time/stop, review and submit inputs.
- FIG. 103 illustrates an annotated view of the menu bar of the artist display.
- the menu bar is shown at the top left portion of the display in Fig. 102.
- FIG. 104 illustrates an annotated view of the task row of the artist display. This annotated view shows only one row, however multiple rows may be displayed as per Fig. 102.
- FIG. 105 illustrates an annotated view of the main portion of the user interface of the artist display of Fig 102.
- FIG. 106 illustrates a build timeline display for the artist display for creating a timeline to work on.
- FIG. 107 illustrates a browse snapshots display for the artist display that enables an artist to view the snapshot of a shot or otherwise cache important information related to a shot so that the database does not have to field requests for often utilized data.
- the snapshot keeps track of the locations of various files associated with a shot, and keeps track of other information related to a work product related to the shot, i.e., source, masks, resolution, file type.
- the snapshot keeps track of versioning of the various files and file types and optionally the versions of the tools utilized to work on the various files.
- FIG. 108 illustrates an artist actual window showing the actuals of time spent on tasks, for example versus time allocated for the tasks, with dropdown menus for specific timesheets.
- FIG. 109 illustrates a notes display for the artist display that enables artists to enter a note related to the shot.
- FIG. 110 illustrates a check in display for the artist display that enables work to be checked in after work on the shot is complete.
- the computer is generally also configured to present a third display configured to be viewed by an editor, i.e., an editorial worker, that includes an annotation frame configured to accept commentary or drawing or both commentary and drawing on the at least one of said plurality of images associated with the at least one shot.
- One or more embodiments of the computer may be configured to provide the third display configured to be viewed by an editor that includes an annotation overlaid on at least one of the plurality of images. This capability provides information on one display that has generally required three workers to integrate in known systems, and which is novel in and of itself.
- FIG. I l l illustrates a view of an editorial display showing the project, filter inputs, timeline inputs and search results showing the shots in the main portion of the window along with the context of work and assigned worker.
- FIG. 112 illustrates a view of the session manager display of the editorial display for selecting shots to review.
- FIG. 113 illustrates a view of the advanced search display of the editorial display.
- FIG. 114 illustrates a view of the simple search display of the editorial display.
- FIG. 115 illustrates a view of the reviewing pane for a shot also showing integrated notes and/or snapshot information in the same frame. This view has in the past generally required three workers to create and saves great amounts of time and speeds the review process greatly. Any type of information may be overlaid onto an image to enable consolidated disparate views of images and related data on one display.
- FIG. 116 illustrates a view of the timelines to select for review and/or checkin after modification.
- FIG. 117 illustrates the annotations added to a frame for feedback using the tool of FIG. 117.
- One or more embodiments of the computer are configured to accept a rating input from production or editorial based on work performed by the artist, optionally in a blind manner in which the reviewer does not know the identity of the artist in order to prevent favoritism for example.
- One or more embodiments of the computer are configured to accept a difficulty of the at least one shot and calculate a rating based on work performed by the artist and based on the difficulty of the shot and time spent on the shot.
- One or more embodiments of the computer are configured to accept a rating input from production or editorial based on work performed by the artist, or, accept a difficulty of the at least one shot and calculate a rating based on work performed by the artist and based on the difficulty of the shot and time spent on the shot, and, signify an incentive with respect to the artist based on the rating accepted by the computer or calculated by the computer.
- One or more embodiments of the computer are configured to estimate remaining cost based on the actuals that are based on total time spent for all of the at least one tasks associated with all of the at least one shot in the project versus time allocated for all of the at least one tasks associated with all of the at least one shot in the project.
- One or more embodiments of the computer are configured to compare the actuals associated with a first project with actuals associated with a second project and signify at least one worker to be assigned from the first project to the second project based on at least one rating of the first worker that is assigned to the first project.
- One or more embodiments of the computer are configured to analyze a prospective project having a number of shots and estimated difficulty per shot and based on actuals associated with the project, calculate a predicted cost for the prospective project.
- One or more embodiments of the computer are configured to analyze a prospective project having a number of shots and estimated difficulty per shot and based on the actuals associated with a first previously performed project and a second previously performed project that completed after the first previously performed project, calculate a derivate of the actuals, calculate a predicted cost for the prospective project based on the derivative of the actuals. For example, as the process improves, tools improve and workers improve, the efficiency of work improves and the budgeting and bid processes can take this into account by calculating how efficiency is changing versus time and use this rate of change to predict costs for a prospective project.
- One or more embodiments of the computer are configured to analyze the actuals associated with said project and divide completed shots by total shots associated with said project and present a time of completion of the project.
- One or more embodiments of the computer are configured to analyze the actuals associated with the project and divide completed shots by total shots associated with the project, present a time of completion of the project, accept an input of at least one additional artist having a rating, accept a number of shots in which to use the additional artist, calculate a time savings based on the at least one additional artist and the number of shots, subtract the time savings from the time of completion of the project and present an updated time of completion of the project.
- One or more embodiments of the computer are configured to calculate amount of disk space that may be utilized to archive the project and signify at least one asset that may be rebuilt from other assets to avoid archival of the at least one asset.
- One or more embodiments of the computer are configured to display an error message if the artist is working with a frame number that is not current in the at least one shot. This may occur when fades, dissolves or other effects lengthen a particular shot for example wherein the shot contains frames not in the original source assets.
- Feature films are tele-cined or transferred from 35mm or 16mm film using a high resolution scanner such as a 10-bit SPIRIT DATACINE ® or similar device to HDTV (1920 by 1080 24P) or data-cined on a laser film scanner such as that manufactured by IMAGICA ® Corp. of America at a larger format 2000 lines to 4000 lines and up to 16bits of grayscale.
- the high resolution frame files are then converted to standard digital files such as uncompressed TIP files or uncompressed TGA files typically in 16bit three-channel linear format or 8bit three channel linear format.
- the 10-bit HDTV frame files are converted to similar TIF or TGA uncompressed files at either 16-bits or 8-bit per channel. Each frame pixel is then averaged such that the three channels are merged to create a single 16bit channel or 8bit channel respectively.
- Any other scanning technologies capable of scanning an existing film to digital format may be utilized.
- many movies are generated entirely in digital format, and thus may be utilized without scanning the movie.
- the metadata can be imported for example to obtain an alpha and/or mask and/or depth for the computer-generated element on a pixel-by-pixel or sub-pixel-by- sub-pixel basis.
- One format of a file that contains alpha/mask and depth data is the RGBAZ file format, of which one implementation is the EXR file format.
- Visual images are transferred from developed motion picture film to a high definition video storage medium, which is a storage medium adapted to store images and to display images in conjunction with display equipment having a scan density substantially greater than that of an NTSC compatible video storage medium and associated display equipment.
- the visual images are also transferred, either from the motion picture film or the high definition video storage medium to a digital data storage format adapted for use with digital nonlinear motion picture editing equipment.
- the digital nonlinear motion picture editing equipment is used to generate an edit decision list, to which the motion picture film is then conformed.
- the high definition video storage medium is generally adapted to store and display visual images having a scan density of at least 1080 horizontal lines.
- Electronic or optical transformation may be utilized to allow use of visual aspect ratios that make full use of the storage formats used in the method.
- This digitized film data as well as data already transferred from film to one of a multiplicity of formats such as HDTV are entered into a conversion system such as the HDTV STILL STORE ® manufactured by AVICA ® Technology Corporation.
- Such large scale digital buffers and data converters are capable of converting digital image to all standard formats such as 1080i HDTV formats such as 720p, and 1080p/24.
- An Asset Management System server provides powerful local and server back ups and archiving to standard SCSI devices, C2-level security, streamlined menu selection and multiple criteria database searches.
- An undersized third pin enters a third perforation spaced along the film from the second pin and then pulls the film obliquely to a reference line extending between the first and second pins to nest against the first and second pins the perforations thereat and register the image frame precisely at the positioning location or aperture.
- a pair of flexible bands extending along the film edges adjacent the positioning location moves progressively into incrementally increasing contact with the film to iron it and clamp its perforations against the gate.
- the pins register the image frame precisely with the positioning location, and the bands maintain the image frame in precise focal position. Positioning can be further enhanced following the precision mechanical capture of images by methods such as that embodied in US Patent No. 4,903,131, Method For The Automatic Correction Of Errors In Image Registration During Film Scanning.
- the digital movie is broken down into scenes and cuts.
- the entire movie is then processed sequentially for the automatic detection of scene changes including dissolves, wipe-a- ways and cuts.
- These transitions are further broken down into camera pans, camera zooms and static scenes representing little or no movement.
- All database references to the above are entered into an edit decision list (EDT) within the database based on standard SMPTE time code or other suitable sequential naming convention.
- Motion is broken down into simple and complex depending on occlusion elements, number of moving objects and quality of the optics (e.g., softness of the elements, etc).
- Files are numbered using sequential SMPTE time code or other sequential naming convention.
- the image files are edited together at 24-frame/sec speed (without field related 3/2 pull down which is used in standard NTSC 30 frame/sec video) onto a DVD using ADOBE ® AFTER EFFECTS ® or similar programs to create a running video with audio of the feature film or TV series. This is used to assist with scene analysis and scene breakdown.
- a database permits the entering of scene, cut, design, key frame and other critical data in time code format as well as descriptive information for each scene and cut.
- Each scene cut is identified relative to camera technique. Time codes for pans, zooms, static backgrounds, static backgrounds with unsteady or drifting camera and unusual camera cuts that require special attention.
- single frames called key frames from each cut of the feature film are selected that contain all the elements within each cut that require color/depth consideration. There may be as many as 1,000 key frames. These frames will contain all the color/depth transform information necessary to apply color/depth to all sequential frames in each cut without additional color choices.
- Historical reference, studio archives and film analysis provides the designer with color references.
- the designer uses an input device such as a mouse, the designer masks features in a selected single frame containing a plurality of pixels and assigns color to them using an HSL color space model based on creative considerations and the grayscale and luminance distribution underlying each mask.
- One or more base colors are selected for image data under each mask and applied to the particular luminance pattern attributes of the selected image feature. Each color selected is applied to an entire masked object or to the designated features within the luminance pattern of the object based on the unique gray-scale values of the feature under the mask.
- a lookup table or color transform for the unique luminance pattern of the object or feature is thus created which represent the color to luminance values applied to the object. Since the color applied to the feature extends the entire range of potential grayscale values from dark to light the designer can insure that as the distribution of the gray-scale values representing the pattern change homogeneously into dark or light regions within subsequent frames of the movie such as with the introduction of shadows or bright light, the color for each feature also remains consistently homogeneous and correctly lighten or darken with the pattern upon which it is applied.
- Depth can imported for computer-generated objects where metadata exists and/or can be assigned to objects and adjusted using embodiments of the invention using an input device such as a mouse to assign objects particular depths including contour depths, e.g., geometric shapes such as an ellipsoid to a face for example.
- contour depths e.g., geometric shapes such as an ellipsoid to a face for example.
- contour depths e.g., geometric shapes such as an ellipsoid to a face for example.
- contour depths e.g., geometric shapes such as an ellipsoid to a face for example.
- the imported depth and/or alpha and/or mask shape can be adjusted if desired. Assigning a fixed distance to foreground objects tends to make the objects appear as cut-outs, i.e., flat. See also FIGS. 42-70.
- the masks representing designed selected color transforms/depth contours in the single design frame are then copied to all subsequent frames in the series of movie frames by one or more methods such as auto-fitting bezier curves to edges, automatic mask fitting based on Fast Fourier Transforms and Gradient Descent Calculation tied to luminance patterns in a subsequent frame relative to the design frame or a successive preceding frames, mask paint to a plurality of successive frames by painting the object within only one frame, auto-fitting vector points to edges and copying and pasting individual masks or a plurality of masks to selected subsequent frames.
- depth information may be "tweened" to account for forward/backward motion or zooming with respect to the camera capture location.
- the alpha and/or mask data is generally correct and may be skipped for reshaping processes since the metadata associated with computer-generated elements is obtained digitally from the original model of an object and hence does not require adjustment in general.
- step 3710 for setting mask fit location to border of CG element to potentially skip large amounts of processing in fitting masks in subsequent frames to reshape the edges to align a photographic element may be morphed or reshaped to provide special effects not originally in a movie scene.
- camera moves are consolidated and separated from motion elements in each scene by the creation of a montage or composite image of the background from a series of successive frames into a single frame containing all background elements for each scene and cut.
- the resulting single frame becomes a representation of the entire common background of a multiplicity of frames in a movie, creating a visual database of all elements and camera offset information within those frames.
- each pixel within a single frame visual database of a background corresponds to an appropriate address within the respective "raw" (unconsolidated) frame from which it was created, any designer determined masking operation and corresponding masking lookup table designation applied to the visual database will be correctly applied to each pixel's appropriate address within the raw film frames that were used to create the single frame composite.
- sets for each scene and cut are each represented by a single frame (the visual database) in which pixels have either single or multiple representations within the series of raw frames from which they were derived. All masking within a single visual database frame will create a one-bit mask per region representation of an appropriate lookup table that corresponds to either common or unique pixel addresses within the sequential frames that created the single composite frame. These address-defined masking pixels are applied to the full resolution frames where total masking is automatically checked and adjusted where necessary using feature, edge detection and pattern recognition routines. Where adjustments are required, i.e., where applied masked region edges do not correspond to the majority of feature edges within the gray scale image, a "red flag" exception comment signals the operator that frame-by-frame adjustments may be necessary.
- the differencing algorithm used for detecting motion objects will generally be able to differentiate dramatic pixel region changes that represent moving objects from frame to frame.
- the resulting masks will be assigned to a default alpha layer that renders that part of the moving object mask transparent.
- an operator using one or more vector or paint tools will designate the demarcation between the moving object and cast shadow.
- the cast shadows will be detected as an extraneous feature relative to the two key motion objects.
- cast shadows are handled by the background lookup table that automatically adjusts color along a luminance scale determined by the spectrum of light and dark gray scale values in the image.
- Action within each frame is isolated via differencing or frame-to-frame subtraction techniques that include vector (both directional and speed) differencing (i.e., where action occurs within a pan) as well as machine vision techniques, which model objects and their behaviors. Difference pixels are then composited as a single frame (or isolated in a tiling mode) representing a multiplicity of frames thus permitting the operator to window regions of interest and otherwise direct image processing operations for computer controlled subsequent frame masking.
- action taking place in multiple frames within a scene can be represented by a single frame visual database in which each unique pixel location undergoes appropriate one bit masking from which corresponding lookup tables are applied.
- the purpose of creating an action composite visual data base is to window or otherwise designate each feature or region of interest that will receive a particular mask and apply region of interest vectors from one key frame element to subsequent key frame elements thus provide operator assistance to the computer processing that will track each region of interest.
- masks are applied to designer designated regions of interest for a single instance of a motion object appearing within the background (i.e., a single frame of action appears within the background or stitched composited background in the proper x, y coordinates within the background corresponding to the single frame of action from which it was derived).
- a single frame of action appears within the background or stitched composited background in the proper x, y coordinates within the background corresponding to the single frame of action from which it was derived.
- an input device such as a mouse the operator uses the following tools in creating the regions of interest for masking.
- projects having associated computer-generated element metadata may import and if necessary, scale the metadata to the units utilized for depth in the project. Since these masks are digitally created, they can be assumed to be accurate throughout the scene and thus the outlines and depths of the computer-generated areas may be ignored for reshaping operations.
- Key Frame Motion Object Creation The operator windows all masked regions of interest on the design frame in succession and directs the computer by various pointing instruments and routines to the corresponding location (regions of interest) on selected key frame motion objects within the visual database thereby reducing the area on which the computer must operate (i.e., the operator creates a vector from the design frame moving object to each subsequent key frame moving object following a close approximation to the center of the region of interest represented within the visual database of the key frame moving object.
- This operator- assisted method restricts the required detection operations that must be performed by the computer in applying masks to the corresponding regions of interest in the raw frames).
- each region of interest within subsequent motion object frames, between two key motion object frames undergoes a computer masking operation.
- the masking operation involves a comparison of the masks in a preceding motion object frame with the new or subsequent Detector File operation and underlying parameters (i.e., mask dimensions, gray scale values and multiple weighting factors that lie within the vector of parameters in the subsequent key frame motion object) in the successive frame.
- This process is aided by the windowing or pointing (using various pointing instruments) and vector application within the visual database.
- the computer will determine a match and will attempt a best fit.
- the uncompressed, high resolution images all reside at the server level, all subsequent masking operations on the regions of interest are displayed on the compressed composited frame in display memory or on a tiled, compressed frame in display memory so that the operator can determine correct tracking and matching of regions.
- a zoomed region of interest window showing the uncompressed region is displayed on the screen to determine visually the region of interest best fit.
- This high-resolution window is also capable of full motion viewing so that the operator can determine whether the masking operation is accurate in motion.
- FIG. 1 a first embodiment as shown in Figure 1 , a plurality of feature film or television film frames 14 a-n representing a scene or cut in which there is a single instance or perceptive of a background 16 (Figure 3).
- Figure 3 a plurality of feature film or television film frames 14 a-n representing a scene or cut in which there is a single instance or perceptive of a background 16
- Figure 3 shows selected samples of the 120 total frames 14 making up the 5 -second pan.
- FIG 2 an isolated background 16 processed scene from the plurality of frames 14a-n represented in Figure 1 in which all motion elements 18 are removed using various subtraction and differencing techniques.
- the separate frames that created the pan are combined into a visual database in which unique and common pixels from each of the 120 frames 14 composing the original pan are represented in the single composite background image 12 shown in Figure 3.
- the single background image 12 is then used to create a background mask overlay 20 representing designer selected color lookup tables in which dynamic pixel colors automatically compensate or adjust for moving shadows and other changes in luminance. For depth projects, any object in the background may be assigned any depth.
- a variety of tools may be utilized to perform the assignment of depth information to any portion of the background including paint tools, geometric icon based tools that allow setting a contour depth to an object, or text field inputs to allow for numeric inputs.
- the composite background shown in Figure 2 for example may also have a ramp function assigned to allow for a nearer depth to be assigned to the left portion of the scene and a linear increase in depth to the right of the image to be automatically assigned. See also FIGS. 42-70.
- operator assisted and automated operations are used to detect obvious anchor points represented by clear edge detected intersects and other contiguous edges n each frame 14 making up the single composite image 12 and over laid mask 20. These anchor points are also represented within the composite image 12 and are used to aide in the correct assignment of the mark to each frame 14 represented by the single composite image 12.
- Anchor points and objects and/or areas that are clearly defined by closed or nearly closed edges are designed as a single mask area and given a single lookup table. Within those clearly delineated regions polygons are created of which anchor points are dominant points. Where there is no clear edge detected to create a perfectly closed region, polygons are generated using the edge of the applied mask.
- the resulting polygon mesh includes the interior of anchor point dominant regions plus all exterior areas between those regions.
- Pattern parameters created by the distribution of luminance within each polygon are registered in a database for reference when corresponding polygonal addresses of the overlying masks are applied to the appropriate addresses of the frames which were used to create the composite single image 12.
- a representative sample of each motion object (M-Object) 18 in the scene 10 receives a mask overlay that represents designer selected color lookup tables/depth assignments in which dynamic pixel colors automatically compensate or adjust for moving shadows and other changes in luminance as the M-Object 18 moves within the scene 10.
- the representative sample are each considered Key M-Objects 18 that are used to define the underlying patterns, edges, grouped luminance characteristics, etc., within the masked M-Object 18. These characteristics are used to translate the design masks from one Key M-Object 18a to subsequent M-Objects 18b along a defined vector of parameters leading to Key M-Object 18c, each Subsequent M-Object becoming the new Key M-Object in succession as masks are applied.
- Key M-Object 18a may be assigned a depth of 32 feet from the camera capture point while Key M-Object 18c may be assigned a depth of 28 feet from the camera capture point.
- the various depths of the object may be "tweened" between the various depth points to allow for realistic three-dimensional motion to occur within the cut without for example requiring wire frame models of all of the objects in the objects in a frame.
- operator assisted and automated operations are used to detect obvious anchor points represented by clear edge detected intersects and other contiguous edges in each motion object used to create a keyframe.
- Anchor points and specific regions of interest within each motion object that are clearly defined by closed or nearly closed edges are designated as a single mask area and given a single lookup table. Within those clearly delineated regions, polygons are created of which anchor points are dominant points. Where there is no clear edge detected to create a perfectly closed region, polygons are generated using the edge of the applied mask.
- the resulting polygon mesh includes the interior of the anchor point dominant regions plus all exterior areas between those regions.
- Pattern parameters created by the distribution of luminance values within each polygon are registered in a database for reference when corresponding polygonal addresses of the overlying masks are applied to the appropriate addresses of the frames that were used to create the composite single frame 12.
- Subsequent or in-between motion key frame objects 18 are processed sequentially.
- the group of masks comprising the motion key frame object remains in its correct address location in the subsequent frame 14 or in the subsequent instance of the next motion object 18.
- the mask is shown as an opaque or transparent color.
- An operator indicates each mask in succession with a mouse or other pointing device and along with its corresponding location in the subsequent frame and/or instance of the motion object.
- the computer uses the prior anchor point and corresponding polygons representing both underlying luminance texture and mask edges to create a best fit to the subsequent instance of the motion object.
- the applied depths of the various objects determine the amount of horizontal translation applied when generating left and right viewpoints as utilized in three-dimensional viewing as one skilled in the art will appreciate.
- the desired objects may be dynamically displayed while shifting by an operator set and observe a realistic depth.
- the depth value of an object determines the horizontal shift applied as one skilled in the art will recognize and which is taught in at least USPN 6,031,564, to Ma et al., the specification of which is hereby incorporated herein by reference.
- Display A key frame that includes all motion objects for that frame is fully masked and loaded into the display buffer along with a plurality of subsequent frames in thumbnail format; typically 2 seconds or 48 frames.
- Figures 5 A and 5B show a series of sequential frames 14a-n loaded into display memory in which one frame 14 is fully masked with the background (key frame) and ready for mask propagation to the subsequent frames 14 via automatic mask fitting methods.
- All frames 14 along with associated masks and/or applied color trans forms/depth enhancements can also be displayed sequentially in real-time (24 frames/sec) using a second (child) window to determine if the automatic masking operations are working correctly.
- stereoscopic glasses or red/blue anaglyph glasses may be utilized to view both viewpoints corresponding to each eye. Any type of depth viewing technology may be utilized to view depth enhanced images including video displays that require no stereoscopic glasses yet which utilizes more than two image pairs which may be created utilizing embodiments of the invention.
- Figures 6A and 6B show the child window displaying an enlarged and scalable single image of the series of sequential images in display memory.
- the Child window enables the operator to manipulate masks interactively on a single frame or in multiple frames during real time or slowed motion.
- Mask Modification Masks can be copied to all or selected frames and automatically modified in thumbnail view or in the preview window. In the preview window mask modification takes place on either individual frames in the display or on multiple frames during real-time motion.
- a single mask (flesh) is propagated automatically to all frames 14 in the display memory.
- the operator could designate selective frames to apply the selected mask or indicate that it is applied to all frames 14.
- the mask is a duplication of the initial mask in the first fully masked frame. Modifications of that mask occur only after they have been propagated.
- Masks are adjusted to compensate for object motion in subsequent frames using various tools based on luminance, pattern and edge characteristics of the image.
- the associated masks are digitally created and can be assumed to be accurate throughout the scene and thus the outlines and depths of the computer-generated areas may be ignored for automatic mask fitting or reshaping operations. Elements that border these objects, may thus be more accurately reshaped since the outlines of the computer-generated elements are taken as correct.
- the shape of the mask at the junction can be taken to be accurate even though there is no visual difference at the junction.
- the computer-generated element mask can be utilized to define the border of the operator-defined mask as per step 3710 of Fig. 37C. This saves processing time since automatic mask fitting in a scene with numerous computer-generated element masks can be minimized.
- the method for automatically modifying both the location and correctly fitting all masks in an image to compensate for movement of the corresponding image data between frames involves the following: [00312] Set Reference Frame Mask and Corresponding Image Data:
- a reference frame (frame 1) is masked by an operator using a variety of means such as paint and polygon tools so that all regions of interest (i.e., features) are tightly covered.
- FFT Fast Fourier Transform
- the bounding box is augmented by an additional margin around the region to accommodate other motion and shape morphing effects.
- the "small dark" pixels shown in Figure 9B are used to calculate a weighed index using bilinear interpolation.
- the grid spacing is currently set at 10 pixels, so that essentially no more than 1 in 50 pixels are presently fit with a gradient descent search. This grid spacing could be a user controllable parameter.
- FFT Fast Fourier Transform
- Fit Value Gradient The displacement vector data derived from the FFT creates a search fit location, and the error surface calculation begins at that offset position, proceeding down (against) the gradient of the error surface to a local minimum of the surface, which is assumed to be the best fit
- This method finds best fit for each next frame pixel or groups of pixels based on the previous frame, using normalized squared differences, for instance in a 10x10 box and finding a minimum down the mean squared difference gradients. This technique is similar to a cross correlation but with a restricted sampling box for the calculation. In this way the corresponding fit pixel in the previous frame can be checked for its mask index, and the resulting assignment is complete.
- Figures 11 A, B and C show a second search box derived from a descent down the error surface gradient (evaluated separately), for which the evaluated error function is reduced, possibly minimized, with respect to the original reference box (evident from visual comparison of the boxes with the reference box in Figures 10A, B, C and D.).
- the error surface gradient is calculated as per definition of the gradient. Vertical and horizontal error deviations are evaluated at four positions near the search box center position, and combined to provide an estimate of the error gradient for that position. The gradient component evaluation is explained with the help of Figure 12.
- Previous vs. Propagated Reference Images The reference image utilized for mask fitting is usually an adjacent frame in a film image-frame sequence. However, it is sometimes preferable to use an adequately fit mask as a reference image (e.g. a key frame mask, or the source frame from which mask regions were propagated/copied).
- the present embodiment provides a switch to disable "adjacent" reference frames, using the propagated masks of the reference image if that frame is defined by a recent propagation event.
- the process of mask fitting In the present embodiment the operator loads n frames into the display buffer. One frame includes the masks that are to be propagated and fitted to all other frames. All or some of the mask(s) are then propagated to all frames in the display buffer. Since the mask-fitting algorithm references the preceding frame or the first frame in the series for fitting masks to the subsequent frame, the first frame masks and/or preceding masks must be tightly applied to the objects and/or regions of interest. If this is not done, mask errors will accumulate and mask fitting will break down. The operator displays the subsequent frame, adjusts the sampling radius of the fit and executes a command to calculate mask fitting for the entire frame. The execution command can be a keystroke or mouse-hotkey command.
- Figure 14 shows that by using the automatic mask fitting routine, the mask data adjusts to the image data by referencing the underlying image data in the preceding image.
- the mask data in later images within the sequence show marked discrepancy relative to the underlying image data. Eye makeup, lipstick, blush, hair, face, dress and hand image data are all displaced relative to the mask data.
- the mask data is adjusted automatically based on the underlying image data from the previous mask and underlying image data. In this Figure 13, the mask data is shown with random colors to show the regions that were adjusted automatically based on underlying pattern and luminance data. The blush and eye makeup did not have edge data to reference and were auto-adjusted on the basis of luminance and grayscale pattern.
- mask data from Figure 16 is shown with appropriate color transforms after whole frame automatic mask fitting.
- the mask data is adjusted to fit the underlying luminance pattern based on data from the previous frame or from the initial key frame.
- Masks for motion objects can be animated using either Bezier curves or polygons that enclose a region of interest.
- a plurality of frames are loaded into display memory and either Bezier points and curves or polygon points are applied close to the region of interest where the points automatically snap to edges detected within the image data.
- the operator adjusts the polygon or Bezier in the last frame of the frames loaded in display memory.
- the operator executes a fitting routine, which snaps the polygons or Bezier points plus control curves to all intermediate frames, animating the mask over all frames in display memory.
- the polygon and Bezier algorithms include control points for rotation, scaling and move-all to handle camera zooms, pans and complex camera moves.
- polygons are used to outline a region of interest for masking in frame one.
- the square polygon points snap to the edges of the object of interest.
- the Bezier points snap to the object of interest and the control points/curves shape to the edges.
- the entire polygon or Bezier curve is carried to a selected last frame in the display memory where the operator adjusts the polygon points or Bezier points and curves using the snap function, which automatically snaps the points and curves to the edges of the object of interest.
- FIG 21 shows the resulting masks from a polygon or Bezier animation with automatic point and curve snap to edges.
- the brown masks are the color transforms and the green masks are the arbitrary color masks.
- areas that have been depth assigned may be of one color while those areas that have yet to be depth assigned may be of another color for example.
- Embodiments of the invention include a method and process for automatically colorizing/depth enhancing a plurality of frames in scenes cuts that include complex camera movements as well as scene cuts where there is camera weave or drifting cameras movement that follows erratic action of the motion objects.
- Camera Pans For a pan camera sequence, the background associated with non-moving objects in a scene form a large part of the sequence. In order to colorize/depth enhance a large amount of background objects for a pan sequence, a mosaic that includes the background objects for an entire pan sequence with moving objects removed is created. This task is accomplished with a pan background stitcher tool. Once a background mosaic of the pan sequence is generated, it can be colorized/depth enhanced once and applied to the individual frames automatically, without having to manually colorize/depth assign the background objects in each frame of the sequence.
- the pan background stitcher tool generates a background image of a pan sequence using two general operations.
- First, the movement of the camera is estimated by calculating the transformation needed to align each frame in the sequence with the previous frame. Since moving objects form a large portion of cinematic sequences, techniques are used that minimize the effects of moving objects on the frame registration.
- Second, the frames are blended into a final background mosaic by interactively selecting two pass blending regions that effectively remove moving objects from the final mosaic.
- Background composite output data includes a greyscale/(or possibly color for depth projects) image file of standard digital format such as TIFF image file (bkg.*.tif) comprised of a background image of the entire pan shot, with the desired moving objects removed, ready for color design/depth assignments using the masking operations already described, and an associated background text data file needed for background mask extraction after associated background mask/colorization/depth data components (bkg.*.msk, bkg.*lut, ...) have been established.
- TIFF image file bkg.*.tif
- the background text data file provides filename, frame position within the mosaic, and other frame-dimensioning information for each constituent (input) frame associated with the background, with the following per line (per frame) content: Frame-filename, frame-x-position, frame -y-position, frame-width, frame-height, frame-left-margin-x-max, frame -right-margin-x- min.
- Each of the data fields are integers except for the first (frame-filename), which is a string.
- the registration method used in the pan stitcher uses properties of the Fourier transform in order to avoid bias towards moving objects in the scene. Automatic registration of frame pairs is calculated and used for the final background image assembly.
- Phase Shift Calculation The next step involves calculating the phase shift between the images. Doing this results in an expression for the phase shift in terms of the Fourier transform of the first and second image:
- Peak Location The two-dimensional surface that results from (4) will have a maximum peak at the translation point from the first image to the second image. By searching for the largest value in the surface, it is simple to find the transform that represents the camera movement in the scene. Although there will be spikes present due to moving objects, the dominant motion of the camera should represent the largest peak value. This calculation is performed for every consecutive pair of frames in the entire pan sequence.
- a background image buffer is generated which is large enough to span the entire sequence.
- the background can be blended together in a single pass, or if moving objects need to be removed, a two-pass blend is used, which is detailed below.
- the position and width of the blend can be edited in the stitching application and can be set globally set or individually set for each frame pair. Each blend is accumulated into the final mosaic and then written out as a single image file.
- Two Pass Blending The objective in two-pass blending is to eliminate moving objects from the final blended mosaic. This can be done by first blending the frames so the moving object is completely removed from the left side of the background mosaic. An example is shown in Figure 23, where the character can is removed from the scene, but can still be seen in the right side of the background mosaic. Figure 23. In the first pass blend shown in Figure 23, the moving character is shown on the stairs to the right [00380] A second background mosaic is then generated, where the blend position and width is used so that the moving object is removed from the right side of the final background mosaic. An example of this is shown in Figure 24, where the character can is removed from the scene, but can still be seen the left side of the background mosaic. In the second pass blend as shown in
- the stitcher application has on option to interactively set the blending width and position for each pass and each frame individually or globally.
- An example screen shot from the blend editing tool, showing the first and second pass blend positions and widths, can be seen in Figure 26, which is a screen shot of the blend-editing tool.
- each text data record includes: Frame-filename frame-x-position frame -y-position frame-width frame-height frame-left-margin-x-max frame -right-margin-x-min.
- the output text data filename is composed from the first composite input frame rootname by prepending the "bkg.” prefix and appending the ".txt” extension.
- Example Representative lines output text data file called "bkgA.00233.txt" that may include data from 300 or more frames making up the blended image. :
- Image offset information used to create the composite representation of the series of frames is contained within a text file associated with the composite image and used to apply the single composite mask to all the frames used to create the composite image.
- Figure 29 show the composite or montage image of the entire camera pan that was built using phase correlation techniques.
- the motion object (butler) included as a transparency for reference by keeping the first and last frame and averaging the phase correlation in two directions.
- the single montage representation of the pan is color designed using the same color transform masking techniques as used for the foreground object.
- Figure 30 shows that the sequence of frames in the camera pan after the background mask color transforms the montage has been applied to each frame used to create the montage.
- the mask is applied where there is no pre-existing mask thus retaining the motion object mask and color transform information while applying the background information with appropriate offsets.
- the left and right eye views of each frame may be shown as pairs, or in a separate window for each eye for example.
- the images may be displayed on a three-dimensional viewing display as well.
- Static and drifting camera shots Objects which are not moving and changing in a film scene cut can be considered “background” objects, as opposed to moving "foreground” objects. If a camera is not moving throughout a sequence of frames, associated background objects appear to be static for the sequence duration, and can be masked and colorized only once for all associated frames. This is the "static camera” (or “static background”) opposed to the moving (e.g. panning) camera case, which requires stitching tool described above to generate a background composite.
- Cuts or frame sequences involving little or no camera motion provide the simplest case for generating frame-image background "composites" useful for cut background colorization.
- the static background composition tool cannot assume perfect pixel alignment from frame-to-frame, requiring an assessment of inter-frame shifts, accurate to 1 pixel, in order to optimally associated pixels between frames prior to adding their data contribution into the composite (an averaged value).
- the Static Background Composite tool provides this capability, generating all the data necessary to later colorize and extract background colorization information for each of the associated frames.
- Moving foreground objects such as actors, etc.
- the masked moving object exposes the background or foreground the instance of background and foreground previously occluded is copied into the single image with priority and proper offsets to compensate for movement.
- the offset information is included in a text file associated with the single representation of the background so that the resulting mask information can be applied to each frame in the scene cut with proper mask offsets.
- Background composite output data uses a greyscale TIFF image file (bkg.*.tif) that includes averaged input background pixel values lending itself to colorization/depth enhancement, and an associated background text data file required for background mask extraction after associated background mask/colorization data/depth enhancement components (bkg.*.msk, bkg.*.lut,...) have been established.
- Background text data provides filename, mask- offset, and other frame-dimensioning information for each constituent (input) frame associated with the composite, with the following per line (per frame) format: Frame-filename frame-x- offset frame-y-offset frame-width frame-height frame-left-margin-x-max frame-right-margin-x- min.
- Each of these data fields are integers except for the first (frame-filename), which is a string.
- Initialization of the static background composition process involves initializing and acquiring the data necessary to create the composited background image-buffer and -data. This requires a loop over all constituent input image frames. Before any composite data initialization can occur, the composite input frames must be identified, loaded, and have all foreground objects identified/colorized (i.e. tagged with mask labels, for exclusion from composite). These steps are not part of the static background composition procedure, but occur prior to invoking the composite tool after browsing a database or directory tree, selecting and loading relevant input frames, painting/depth assigning the foreground objects.
- Adjacent frames' image background data in a static camera cut may exhibit small mutual vertical and horizontal offsets. Taking the first frame in the sequence as a baseline, all successive frames' background images are compared to the first frames', fitting line- wise and column-wise, to generate two histograms of "measured" horizontal and vertical offsets, from all measurable image-lines and -columns. The modes of these histograms provide the most frequent (and likely) assessed frame offsets, identified and stored in arrays DVx[iframe], DVy [iframe] per frame [iframe]. These offset arrays are generated in a loop over all input frames.
- Fig. 39A shows the edgeDetection and determination of points to snap to (1.1 and 1.2 respectively), which are detailed in Figures 39B and 39C respectively and which enable one skilled in the art to implement a image edge detection routine via Average Filter, Gradient Filter, Fill Gradient Image and a comparison with a Threshold.
- the GetSnapPoint routine of Fig. 39C shows the determination of a NewPoint based on the BestSnapPoint as determined by the Rangelmage less than MinDistance as shown.
- Figs. 40A-C shows how a bimodal threshold tool is implemented in one or more embodiments of the invention. Creation of an image of light and dark cursor shape is implemented with the MakeLightShape routine wherein the light/dark values for the shape are applied with the respective routine as shown at the end of Fig. 40A. These routines are shown in Fig. 40C and 40B respectively. Figs. 41A-B show the calculation of FitValues and gradients for use in one or more of the above routines.
- Composite Frame Loop Input frames are composited (added) sequentially into the resultant background via a loop over the frames. Input frame background pixels are added into the background image buffer with the relevant offset (DVx[iframe], DVy[iframe]) for each frame, and associated pixel composite count values are incremented by one (1) for pixels receiving a composite addition (a separate composite count array/buffer is provided for this). Only background pixels, those without an associated input mask index, are composited (added) into the resultant background; pixels with nonzero (labeled) mask values are treated as foreground pixels and are therefore not subject to composition into the background; thus they are ignored. A status bar in the Gill is incremented per pass through the input frame loop.
- Composite Finish The final step in generating the output composite image buffer requires evaluating pixel averages which constitute the composite image. Upon completion of the composite frame loop, a background image pixel value represents the sum of all contributing aligned input frame pixels. Since resultant output pixels must be an average of these, division by a count of contributing input pixels is required. The count per pixel is provided by the associated pixel composite count buffer, as mentioned. All pixels with nonzero composite counts are averaged; other pixels remain zero.
- Composite Image Save A TIFF format output gray-scale image with 16 bits per pixel is generated from composite-averaged background image buffer.
- the output filename is composed from the first composite input frame filename by pre -pending the "bkg.” prefix (and appending the usual ".tif image extension if required), and writing to the associated background folder at path "../Bckgrnd Frm", if available, otherwise to the default path (same as input frames').
- the output text data filename is composed from the first composite input frame rootname by prepending the "bkg.” prefix and appending the ".txt” extension, and writing to the associated background folder at path "../Bckgrnd Frm", if available, otherwise to the default path (same as input frames').
- Data Cleanup Releases memory allocated to data objects used by the static background composite procedure. These include the background composite GUI dialog object and its member arrays DVx[], DVy[], lMarg[], rMarg[], and the background composite image buffer object, whose contents have previously been saved to disk and are no longer needed.
- the offset data for the background composite is transferred to the mask data overlaying the background such that the mask for each successive frame used to create the composite is placed appropriately.
- the background mask data is applied to each successive frame wherever there are no pre-existing masks (e.g. the foreground actors).
- Figure 32 shows a sequence of frames in which all moving objects (actors) are masked with separate color transforms/depth enhancements.
- Figure 33 shows a sequence of selected frames for clarity prior to background mask information. All motion elements have been fully masked using the automatic mask-fitting algorithm.
- Figure 34 shows the stationary background and foreground information minus the previously masked moving objects.
- the single representation of the complete background has been masked with color transforms in a manner similar to the motion objects.
- outlines of removed foreground objects appear truncated and unrecognizable due to their motion across the input frame sequence interval, i.e., the black objects in the frame represent areas in which the motion objects (actors in this case) never expose the background and foreground, i.e., missing background image data 3401.
- the black objects are ignored for colorization-only projects during the masking operation because the resulting background mask is later applied to all frames used to create the single representation of the background only where there is no pre-existing mask.
- the black objects where missing background image data 3401 exists may artistically or realistically rendered, for example to fill in information to be utilized in the conversion of two-dimensional images into three-dimensional images. Since these areas are areas where pixels may not be borrowed from other frames since they are never exposed in a scene, drawing them or otherwise creating believable images there, allows for all background information to be present and used for artifact free two-dimensional to three-dimensional conversion. For example, in order to create artifact-free three-dimensional image pairs from a two-dimensional image having areas that are never exposed in a scene, backgrounds having all or enough required information for the background areas that are always occluded may be generated.
- the missing background image data 3401 may be painted, drawn, created, computer-generated or otherwise obtained from a studio for example, so that there is enough information in a background, including the black areas to translate foreground objects horizontally and borrow generated background data for the translated edges for occluded areas.
- Figure 35 shows the sequential frames in the static camera scene cut after the background mask information has been applied to each frame with appropriate offsets and where there is no pre-existing mask information.
- Figure 36 shows a representative sample of frames from the static camera scene cut after the background information has been applied with appropriate offsets and where there is no pre-existing mask information.
- Colorization Rendering After color processing is completed for each scene, subsequent or sequential color motion masks and related lookup tables are combined within 24-bit or 48-bit RGB color space and rendered as TIF or TGA files. These uncompressed, high-resolution images are then rendered to various media such as HDTV, 35mm negative film (via digital film scanner), or a variety of other standard and non standard video and film formats for viewing and exhibit.
- 35mm film is digitized to 1920x1080x10 in any one of several digital formats.
- Each frame undergoes standard stabilization techniques to minimize natural weaving motion inherent in film as it traverses camera sprockets as well as any appropriate digital telecine technology employed. Frame-differencing techniques are also employed to further stabilize image flow.
- Background and foreground elements may include computer-generated elements or elements that exist in the original movie footage for example.
- Foregrounds are defined as any object and/or region that move in the same direction as the background but may represent a faster vector because of its proximity to the camera lens. In this method pans are reduced to a single representative image, which contains all of the background and foreground information taken from a plurality of frames.
- Zooms are sometimes handled as a tiled database in which a matrix is applied to key frames where vector points of reference correspond to feature points in the image and correspond to feature points on the applied mask on the composited mask encompassing any distortion.
- a database is created from the frames making up the single representative or composited frame (i.e., each common and novel pixel during a pan is assigned to the plurality of frames from which they were derived or which they have in common).
- Each entire background is colorized/depth assigned as a single frame in which all motion objects are removed.
- Background masking is accomplished using a routine that employs standard paint, fill, digital airbrushing, transparency, texture mapping, and similar tools.
- Color selection is accomplished using a 24-bit color lookup table automatically adjusted to match the density of the underlying gray scale and luminance.
- Depth assignment is accomplished via assigning depths, assigning geometric shapes, entry of numeric values with respect to objects, or in any other manner in the single composite frame. In this way creatively selected colors/depths are applied that are appropriate for mapping to the range of gray scale/depth underlying each mask.
- the standard color wheel used to select color ranges detects the underlying grayscale dynamic range and determines the corresponding color range from which the designer may choose (i.e., only from those color saturations that will match the grayscale luminance underlying the mask.)
- Each lookup table allows for a multiplicity of colors applied to the range of gray scale values underlying the mask.
- the assigned colors will automatically adjust according to luminance and/or according to pre-selected color vectors compensating for changes in the underlying gray scale density and luminance.
- Design motion object frames are created which include the entire scene background as well as a single representative moment of movement within the scene in which all characters and elements within the scene are present. These moving non-background elements are called Design Frame Objects (DFO).
- DFO Design Frame Objects
- Each DFO is broken down into design regions of interest (regions of interest) with special attention focused on contrasting elements within the DFOs that can be readily be isolated using various gray scale and luminance analyses such as pattern recognition and or edge detection routines. As existing color movies may be utilized for depth enhancement, regions of interest may be picked with color taken into account.
- Color selection is determined for each region of interest comprising each object based on appropriate research into the film genre, period, creative intention, etc. and using a 24bit color lookup table automatically adjusted to match the density of the underlying gray scale and luminance suitable and creatively selected colors are applied.
- the standard color wheel detects the underlying grayscale range and restricts the designer to choose only from those color saturations that will match the grayscale luminance underlying the mask. Depth assignments may be made or adjusted for depth projects until realistic depth is obtained for example.
- Pre-Production Design Key Frame Objects Assistant Designer [00455] 1. Once all color selection/depth assignment is generally completed for a particular scene the design motion object frame is then used as a reference to create the larger number of key frame objects within the scene.
- the determining factor for each successive key frame object is the amount of new information between one key frame and the next key frame object.
- the Production Colorist loads a plurality of frames into the display buffer.
- One of the frames in the display buffer will include a key frame from which the operator obtains all masking information. The operator makes no creative or color/depth decisions since all color transform information is encoded within the key frame masks.
- the operator can view the motion of all frames in the display buffer observing the motion that occurs in successive frames or they can step through the motion from one key frame to the next.
- Figs. 37A shows the mask fitting generally processing flow chart that is broken into subsequent detailed flow charts 37B and 37C.
- the program makes a best fit based on the grayscale/luminance, edge parameters and pattern recognition based on the gray scale and luminance pattern of the key frame or the previous frame in the display.
- the mask fitting routines are skipped since the masks or alphas define digitally created (and hence non-operator-defined) edges that accurately define the computer-generated element boundaries.
- Mask fitting operations take into account the computer-generated element masks or alphas and stop when hitting the edge of a computer-generated element mask since these boundaries are accepted as accurate irrespective of grey-scale as per step 3710 of Fig. 37C. This enhances the accuracy of mask edges and reshapes when colors of a computer-generated element and operator-defined mask are of the same base luminance for example.
- the Mask Fit initializes the region and fit grid parameters, then calls the Calculate fit grid routine and then the Interpolate mask on fit grid routine, which execute on any computer as described herein, wherein the routines are specifically configured to calculate fit grids as specified in Figs. 37B and 37C.
- the indices i and j for the FitGridCell are determined and a bilinear interpolation is performed at the fitGridA-D locations wherein the Mask is fit up to any border found for any CG element at 3710 (i.e., for a known alpha border or border with depth values for example that define a digitally rendered element that is taken as a certified correct mask border).
- the mask fitting interpolation is continued up to the size of the mask defined by xend and yend.
- Masks for motion objects can also be animated using either Bezier curves or polygons that enclose a region of interest.
- a plurality of frames are loaded into display memory and either Bezier points and curves of polygon points are applied close to the region of interest where the points automatically snap to edges detected within the image data.
- the polygon and Bezier algorithms include control points for rotation, scaling and move-all to handle zooms, pans and complex camera moves where necessary.
- FIG. 42 shows two image frames that are separated in time by several frames, of a person levitating a crystal ball wherein the various objects in the image frames are to be converted from two-dimensional objects to three-dimensional objects.
- the crystal ball moves with respect to the first frame (shown on top) by the time that the second frame (shown on the bottom) occurs.
- the frames are associated with one another, although separated in time, much of the masking information can be utilized for both frames, as reshaped using embodiments of the invention previously described above.
- FIG. 43 shows the masking of the first object in the first image frame that is to be converted from a two-dimensional image to a three-dimensional image.
- the first object masked is the crystal ball.
- a simple free form drawing tool is utilized to apply a somewhat round mask to the crystal ball.
- a circle mask may be dropped on the image and resized and translated to the correct position to correspond to the round crystal ball.
- the alternative approach is shown herein. The grey-scale values of the masked object are thus utilized to reshape the mask in subsequent frames.
- FIG. 44 shows the masking of the second object in the first image frame.
- the hair and face of the person behind the crystal ball are masked as the second object using a free form drawing tool.
- Edge detection or grey-scale thresholds can be utilized to accurately set the edges of the masks as has been previously described above with respect to colorization.
- an object be a single object, i.e., the hair and face of a person can be masked as a single item, or not and depth can thus be assigned to both or individually as desired.
- FIG. 45 shows the two masks in color in the first image frame allowing for the portions associated with the masks to be viewed. This figure shows the masks as colored transparent masks so that the masks can be adjusted if desired.
- FIG. 46 shows the masking of the third object in the first image frame.
- the hand is chosen as the third object.
- a free form tool is utilized to define the shape of the mask.
- FIG. 47 shows the three masks in color in the first image frame allowing for the portions associated with the masks to be viewed. Again, the masks can be adjusted if desired based on the transparent masks.
- FIG. 48 shows the masking of the fourth object in the first image frame. As shown the person's jacket form the fourth object.
- FIG. 49 shows the masking of the fifth object in the first image frame. As shown the person's sleeve forms the fifth object.
- FIG. 50 shows a control panel for the creation of three-dimensional images, including the association of layers and three-dimensional objects to masks within an image frame, specifically showing the creation of a Plane layer for the sleeve of the person in the image.
- the "Rotate” button is enabled, shown a "Translate Z” rotation quantity showing that the sleeve is rotated forward as is shown in the next figure.
- FIG. 51 shows a three-dimensional view of the various masks shown in FIGS. 43-49, wherein the mask associated with the sleeve of the person is shown as a Plane layer that is rotated toward the left and right viewpoints on the right of the page. Also, as is shown the masks associated with the jacket and person's face have been assigned a Z-dimension or depth that is in front of the background.
- FIG. 52 shows a slightly rotated view of FIG. 51. This figure shows the Plane layer with the rotated sleeve tilted toward the viewpoints. The crystal ball is shown as a flat object, still in two-dimensions as it has not yet been assigned a three-dimensional object type.
- FIG. 53 shows a slightly rotated view of FIG. 51 (and 52), wherein the sleeve is shown tilting forward, again without ever defining a wire frame model for the sleeve.
- a three-dimensional object type of column can be applied to the sleeve to make an even more realistically three-dimensional shaped object.
- the Plane type is shown here for brevity.
- FIG. 54 shows a control panel specifically showing the creation of a sphere object for the crystal ball in front of the person in the image.
- the Sphere three-dimensional object is created and dropped into the three-dimensional image by clicking the "create selected” button in the middle of the frame, which is then shown (after translation and resizing onto the crystal ball in the next figure).
- FIG. 55 shows the application of the sphere object to the flat mask of the crystal ball, that is shown within the sphere and as projected to the front and back of the sphere to show the depth assigned to the crystal ball.
- the Sphere object can be translated, i.e., moved in three axis, and resized to fit the object that it is associated with.
- the projection of the crystal ball onto the sphere shows that the Sphere object is slightly larger than the crystal ball, however this ensures that the full crystal ball pixels are assigned depths.
- the Sphere object can be resized to the actual size of the sphere as well for more refined work projects as desired.
- FIG. 56 shows a top view of the three-dimensional representation of the first image frame showing the Z-dimension assigned to the crystal ball shows that the crystal ball is in front of the person in the scene.
- FIG. 57 shows that the sleeve plane rotating in the X-axis to make the sleeve appear to be coming out of the image more.
- the circle with a line (X axis line) projecting through it defines the plane of rotation of the three-dimensional object, here a plane associated with the sleeve mask.
- FIG. 58 shows a control panel specifically showing the creation of a Head object for application to the person's face in the image, i.e., to give the person's face realistic depth without requiring a wire model for example.
- the Head object is created using the "Created Selected" button in the middle of the screen and is shown in the next figure.
- FIG. 59 shows the Head object in the three-dimensional view, too large and not aligned with the actual person's head.
- the Head object shows up in the three-dimensional view as a generic depth primitive that is applicable to heads in general. This is due to the fact that depth information is not exactly required for the human eye. Hence, in depth assignments, generic depth primitives may be utilized in order to eliminate the need for three-dimensional wire frames.
- the Head object is translated, rotated and resized in subsequent figures as detailed below.
- FIG. 60 shows the Head object in the three-dimensional view, resized to fit the person's face and aligned, e.g., translated to the position of the actual person's head.
- FIG. 61 shows the Head object in the three-dimensional view, with the Y-axis rotation shown by the circle and Y-axis originating from the person's head thus allowing for the correct rotation of the Head object to correspond to the orientation of the person's face.
- FIG. 62 shows the Head object also rotated slightly clockwise, about the Z-axis to correspond to the person's slightly tilted head.
- the mask shows that the face does not have to be exactly lined up for the result three-dimensional image to be believable to the human eye. More exacting rotation and resizing can be utilized where desired.
- FIG. 63 shows the propagation of the masks into the second and final image frame. All of the methods previously disclosed above for moving masks and reshaping them are applied not only to colorization but to depth enhancement as well. Once the masks are propagated into another frame, all frames between the two frames may thus be tweened. By tweening the frames, the depth information (and color information if not a color movie) are thus applied to non-key frames.
- FIG. 64 shows the original position of the mask corresponding to the person's hand.
- FIG. 65 shows the reshaping of the mask, that is performed automatically and with can be adjusted in key frames manually if desired, wherein any intermediate frames get the tweened depth information between the first image frame masks and the second image frame masks.
- the automatic tracking of masks and reshaping of the masks allows for great savings in labor. Allowing manual refinement of the masks allows for precision work where desired.
- FIG. 66 shows the missing information for the left viewpoint as highlighted in color on the left side of the masked objects in the lower image when the foreground object, here a crystal ball is translated to the right.
- highlighted data In generating the left viewpoint of the three-dimensional image, highlighted data must be generated to fill the missing information from that viewpoint.
- FIG. 67 shows the missing information for the right viewpoint as highlighted in color on the right side of the masked objects in the lower image when the foreground object, here a crystal ball is translated to the left.
- the highlighted data must be generated to fill the missing information from that viewpoint.
- a single camera viewpoint may be offset from the viewpoint of the original camera, however the missing data is large for the new viewpoint. This may be utilized if there are a large number of frames and some of the missing information is found in adjacent frames for example.
- FIG. 68 shows an anaglyph of the final depth enhanced first image frame viewable with Red/Blue 3-D glasses. The original two-dimensional image is now shown in three-dimensions.
- FIG. 69 shows an anaglyph of the final depth enhanced second and last image frame viewable with Red/Blue 3-D glasses, note rotation of person's head, movement of person's hand and movement of crystal ball.
- the original two-dimensional image is now shown in three- dimensions as the masks have been moved/reshaped using the mask tracking/reshaping as described above and applying depth information to the masks in this subsequent frame from an image sequence.
- the operations for applying the depth parameter to a subsequent frame is performed using a general purpose computer having a central processing unit (CPU), memory, bus situated between the CPU and memory for example specifically programmed to do so wherein figures herein which show computer screen displays are meant to represent such a computer.
- CPU central processing unit
- FIG. 70 shows the right side of the crystal ball with fill mode "smear", wherein the pixels with missing information for the left viewpoint, i.e., on the right side of the crystal ball are taken from the right edge of the missing image pixels and "smeared" horizontally to cover the missing information.
- Any other method for introducing data into hidden areas is in keeping with the spirit of the invention. Stretching or smearing pixels where missing information is creates artifacts that are recognizable to human observers as errors.
- By obtaining or otherwise creating realistic data for the missing information is, i.e., for example via a generated background with missing information filled in, methods of filling missing data can be avoided and artifacts are thus eliminated.
- providing a composite background or frame with all missing information designated in a way that an artist can use to create a plausible drawing or painting of a missing area is one method of obtaining missing information for use in two-dimensional to three-dimensional conversion projects.
- FIG. 71 shows a mask or alpha plane for a given frame of a scene, for an actor's upper torso and head 7101, and transparent wings 7102.
- the mask may include opaque areas shown as black and transparent areas that are shown as grey areas.
- the alpha plane may be generated for example as an 8 bit grey-scale "OR" of all foreground masks. Any other method of generating a foreground mask having motion objects or foreground object related masks defined is in keeping with the spirit of the invention.
- FIG. 72 shows an occluded area, i.e., missing background image data 7201 as a colored sub-area of the actor of FIG. 71 that never uncovers the underlying background, i.e., where missing information in the background for a scene or frame occurs.
- This area is the area of the background that is never exposed in any frame in a scene and hence cannot be borrowed from another frame.
- any background pixel not covered by a motion object mask or foreground mask can have a simple Boolean TRUE value, all other pixels are thus the occluded pixels as is also shown in FIG. 34.
- FIG. 73 shows the occluded area of FIG. 72 with generated data 7201a for missing background image data that is artistically drawn or otherwise rendered to generate a complete and realistic background for use in artifact free two-dimensional to three-dimensional conversion. See also Fig. 34 and the description thereof. As shown, FIG. 73 also has masks drawn on background objects, which are shown in colors that differ from the source image. This allows for colorization or colorization modifications for example as desired.
- FIG. 73 A shows the occluded area with missing background image data 7201b partially drawn or otherwise rendered to generate just enough of a realistic looking background for use in artifact free two-dimensional to three-dimensional conversion.
- An artist in this example may draw narrower versions of the occluded areas, so that offsets to foreground objects would have enough realistic background to work with when projecting a second view, i.e., translating a foreground object horizontally which exposes occluded areas.
- the edges of the missing background image data area may be drawn horizontally inward by enough to allow for some of the generated data to be used, or all of the generated data to be used in generating a second viewpoint for a three-dimensional image set.
- a number of scenes from a movie may be generated for example by computer drawing by artists or sent to artists for completion of backgrounds.
- a website may be created for artists to bid on background completion projects wherein the website is hosted on a computer system connected for example to the Internet. Any other method for obtaining backgrounds with enough information to render a two-dimensional frame into a three-dimensional pair of viewpoints is in keeping with the spirit of the invention, including rendering a full background with realistic data for all of the occluded area of FIG. 72 (which is shown in FIG. 73) or only a portion of the edges of the occluded area of FIG. 72, (which is shown as FIG. 73A).
- a fixed offset e.g., 100 pixels on each edge of each occluded area, or a percentage of the size of the foreground object, i.e., 5% for example, may flagged to be created and if more data is needed, then the frame is flagged for updating, or smearing or pixel stretching may be utilized to minimize the artifacts of missing data.
- FIG. 74 shows a light area of the shoulder portion on the right side of FIG. 71, where missing background image data 7201 exists when generating a right viewpoint for a right image of a three-dimensional image pair. Missing background image data 7201 represents a gap where stretching (as is also shown in FIG. 70) or other artifact producing techniques would be used when shifting the foreground object to the left to create a right viewpoint.
- the dark portion of the figure is taken from the background where data is available in at least one frame of a scene.
- FIG. 75 shows an example of the stretching of pixels, or "smeared pixels" 7201c, corresponding to the light area in FIG. 74, i.e., missing background image data 7201, wherein the pixels are created without the use of a generated background, i.e., if no background data is available for an area that is occluded in all frames of a scene.
- FIG. 76 shows a result of a right viewpoint without artifacts on the edge of the shoulder of the person through use of generated data 7201a (or 7201b) for missing background image data 7201 shown as for always-occluded areas of a scene.
- FIG. 77 shows an example of a computer-generated element, here robot 7701, which is modeled in three-dimensional space and projected as a two-dimensional image.
- the background is grey to signify invisible areas.
- metadata such as alpha, mask, depth or any combination thereof is utilized to speed the conversion process from two- dimensional image to a pair of two-dimensional images for left and right eye for three- dimensional viewing.
- Masking this character by hand, or even in a computer-aided manner by an operator is extremely time consuming since there are literally hundreds if not thousands of sub- masks required to render depth (and/or color) correctly to this complex object.
- FIG. 78 shows an original image separated into background 7801 and foreground elements 7802 and 7803, (mountain and sky in the background and soldiers in the bottom left also see Fig. 79) along with the imported color and depth of the computer-generated element, i.e., robot 7803 with depth automatically set via the imported depth metadata.
- the soldiers exist in the original image, their depths are set by an operator, and generally shapes or masks with varying depths are applied at these depths with respect to the original objects to obtain a pair of stereo images for left and right eye viewing. (See Fig. 79).
- any area that is covered for the scene such as outline 7804 (of a soldier's head projected onto the background) can be artistically rendered for example to provide believable missing data, as is shown in Fig. 73 based on the missing data of Fig. 73 A, which results in artifact free edges as shown in Fig. 76 for example.
- Importing data for computer generated elements may include reading a file that has depth information on a pixel-by-pixel basis for computer-generated element 7701 and displaying that information in a perspective view on a computer display as an imported element, e.g., robot 7803. This import process saves enormous amounts of operator time and makes conversion of a two-dimensional movie into a three-dimensional movie economically viable.
- One or more embodiments of the invention store the masks and imported data in computer memory and/or computer disk drives for use by one or more computers in the conversion process.
- FIG. 79 shows mask 7901 (forming a portion of the helmet of the rightmost soldier) associated with the photograph of soldiers 7802 in the foreground.
- Mask 7901 along with all other operated-defined masks shown in multiple artificial colors on the soldiers, to apply depth to the various portions of the soldiers occurring in the original image that lie in depth in front of the computer-generated element, i.e., robot 7803.
- the dashed lines horizontally extending from the mask areas 7902 and 7903 show horizontal translation of the foreground objects takes place and where imported metadata can be utilized to accurately auto-correct over-painting of depth or color on the masked objects when metadata exists for the other elements of a movie.
- One type of file that can be utilized to obtain mask edge data is a file with alpha file and/or mask data such as an RGBA file. (See Fig. 80).
- use of generated data for missing areas of the background at these horizontally translated mask areas 7902 and 7903 enables artifact free two-dimensional to three-dimensional conversion.
- FIG. 80 shows an imported alpha layer 8001 shown as a dark blue overlay, which can also be utilized as a mask layer to limit the operator defined, and potentially less accurate masks used for applying depth to the edges of the three soldiers 7802 and designated as soldiers A, B and C.
- an optional computer-generated element such as dust can be inserted into the scene along the line annotated as "DUST", to augment the reality of the scene if desired.
- Any of the background, foreground or computer-generated elements can be utilized to fill portions of the final left and right image pairs as is required.
- FIG. 81 shows the result of using the operator-defined masks without adjustment when overlaying a motion element such as the soldier on the computer-generated element such as the robot.
- a motion element such as the soldier on the computer-generated element such as the robot.
- artifacts occur where operator-defined masks do not exactly align with the edges of the masked objects.
- the soldier's lips show a light colored edge 8101 while the lower picture shows an artifact free edge since the alpha of Figure 80 is used to limit the edges of any operator-defined masks.
- the alpha metadata of FIG. 80 applied to the operated-defined mask edges of FIG. 79, artifact free edges on the overlapping areas is thus enabled.
- application of successively nearer elements combined with their alphas is used to layer all of the objects at their various depths from back to front to create a final image pair for left eye and right eye viewing.
- Embodiments of the invention enable real-time editing of 3D images without re- rendering for example to alter layers/colors/masks and/or remove artifacts and to minimize or eliminate iterative workflow paths back through different workgroups by generating translation files that can be utilized as portable pixel-wise editing files.
- a mask group takes source images and creates masks for items, areas or human recognizable objects in each frame of a sequence of images that make up a movie.
- the depth augmentation group applies depths, and for example shapes, to the masks created by the mask group.
- left and right viewpoint images and left and right translation files may be generated by one or more embodiments of the invention.
- the left and right viewpoint images allow 3D viewing of the original 2D image.
- the translation files specify the pixel offsets for each source pixel in the original 2D image, for example in the form of UV or U maps. These files are generally related to an alpha mask for each layer, for example a layer for an actress, a layer for a door, a layer for a background, etc. These translation files, or maps are passed from the depth augmentation group that renders 3D images, to the quality assurance workgroup.
- the Z depth of regions within the image may also be passed along with the alpha mask to the quality assurance group, who may then adjust depth as well without re-rendering with the original rendering software. This may be performed for example with generated missing background data from any layer so as to allow "downstream" real-time editing without re-rendering or ray-tracing for example.
- Quality assurance may give feedback to the masking group or depth augmentation group for individuals so that these individuals may be instructed to produce work product as desired for the given project, without waiting for, or requiring the upstream groups to rework anything for the current project.
- FIG. 82 shows a source image to be depth enhanced and provided along with left and right translation files (see Figs. 85A-D and 86A-D for embodiments of translation files) and alpha masks (such as shown in Fig. 79) to enable real-time editing of 3D images without re- rendering or ray-tracing the entire image sequence in a scene (e.g., by downstream workgroups) for example to alter layers/colors/masks and/or remove and/or or adjust depths or otherwise change the 3D images without iterative workflow paths back to the original workgroups (as per Fig. 96 versus Fig. 95).
- left and right translation files see Figs. 85A-D and 86A-D for embodiments of translation files
- alpha masks such as shown in Fig. 79
- FIG. 83 shows masks generated by the mask workgroup for the application of depth by the depth augmentation group, wherein the masks are associated with objects, such as for example human recognizable objects in the source image of FIG. 82.
- unskilled labor is utilized to mask human recognizable objects in key frames within a scene or sequence of images.
- the unskilled labor is cheap and generally located offshore. Hundreds of workers may be hired at low prices to perform this tedious work associated with masking.
- Any existing colorization masks may be utilized as a starting point for 3D masks, which may be combined to form a 3D mask outline that is broken into sub-masks that define differing depths within a human recognizable object. Any other method of obtaining masks for areas of an image are in keeping with the spirit of the invention.
- FIG. 84 shows areas where depth is applied generally as darker for nearer objects and lighter for objects that are further away. This view gives a quick overview of the relative depths of objects in a frame.
- FIG. 85A shows a left UV map containing translations or offsets in the horizontal direction for each source pixel.
- translation maps that map the offsets of horizontal movement of individual pixels in a graphical manner may be utilized.
- FIG. 85B shows a right UV map containing translations or offsets in the horizontal direction for each source pixel. Since each of these images looks the same, it is easier to observe that there are subtle differences in the two files by shifting the black value of the color, so as to highlight the differences in a particular area of Figs. 85A and 85B.
- FIG. 85C shows a black value shifted portion of the left UV map of FIG. 85A to show the subtle contents therein.
- FIG. 85D shows a black value shifted portion of the right UV map of FIG. 85B to show the subtle contents therein.
- the branches shown in the slight variances of color signify that those pixels would be shifted to the corresponding location in a pure UV map that maps Red from darkest to lightest in the horizontal direction and maps Green from darkest to lightest in the vertical direction.
- the translation map in the UV embodiment is a graphical depiction of the shifting that occurs when generating a left and right viewpoint with respect to the original source image.
- UV maps may be utilized, however, any other file type that contains horizontal offsets from a source image on a pixel-by-pixel basis (or finer grained) may be utilized, including compressed formats that are not readily viewable as images.
- Some software packages for editing come with pre-built UV widgets, and hence, UV translation files or maps can therefore be utilized if desired.
- certain compositing programs have pre-built objects that enable UV maps to be readily utilized and otherwise manipulated graphically and hence for these implementations, graphically viewable files may be utilized, but are not required.
- FIG. 86A shows a left U map containing translations or offsets in the horizontal direction for each source pixel.
- FIG. 86B shows a right U map containing translations or offsets in the horizontal direction for each source pixel.
- FIG. 86C shows a black value shifted portion of the left U map of FIG. 86A to show the subtle contents therein.
- FIG. 86D shows a black value shifted portion of the right U map of FIG.
- U map or UV map
- Foreground portions of the U map appear darker since they are shifting left and visa versa. This is easy to observe by looking at something in the foreground with only the right eye open and then moving slightly to the right (to observe that the foreground object has indeed been shifted to the left). Since the U map (or UV map) in the unaltered state is a simple ramp of color from dark to light, it then follows that shifting something to the left, i.e., for the right viewpoint, maps it to a darker area of the U map
- each U map (or UV map).
- UV map (or UV map)
- the same tree branches in the same area of each U map (or UV map) are darker for the right eye and brighter for the left eye with respect to un-shifted pixels.
- use of a viewable map is not required, but shows the concept of shifting that occurs for a given viewpoint.
- FIG. 87 shows known uses for UV maps, wherein a three-dimensional model is unfolded so that an image in UV space can be painted onto the 3D model using the UV map.
- This figure shows how UV maps have traditionally been utilized to apply a texture map to a 3D shape.
- the texture here a painting or flat set of captured images of the Earth is mapped to a U and V coordinate system, that is translated to an X, Y and Z coordinate on the 3D model.
- Traditional animation has been performed in this manner in that wire frame models are unraveled and flattened, which defines the U and V coordinate system in which to apply a texture map.
- Embodiments of the invention described herein utilize UV and U maps in a new manner in that a pair of maps are utilized to define the horizontal offsets for two images (left and right) that each source pixel is translated to as opposed to a single map that is utilized to define a coordinate onto which a texture map is placed on a 3D model or wire frame.
- embodiments of the invention utilize UV and U maps (or any other horizontal translation file format) to allow for adjustments to the offset objects without re-rendering the entire scene.
- embodiments of the invention enabled herein utilize two maps, i.e., one for a left and one for a right eye, that map horizontal translations for the left and right viewpoints.
- embodiments of the invention map within one-dimension on a horizontal line -by-line basis.
- the known art maps 2 dimensions to 3 dimensions
- embodiments of the invention utilize 2 maps of translations within 1 dimension (hence visible embodiments of the translation map can utilize one color). For example, if one line of a translation file contains 0, 1, 2, 3...1918, 1919, and the 2 nd and 3 rd pixels are translated right by 4 pixels, then the line of the file would read 0, 4,
- UV files similarly have a ramp function in the Y or vertical axis as well, the values in such a file would be (0,0), (0,1), (0,2)... (0, 1918), (0,1919) corresponding to each pixel, for example for the bottom row of the image and (1,0), (1,1), etc., for the second horizontal line, or row for example.
- This type of offset file allows for movement of pixels in non-horizontal rows, however embodiments of the invention simply shift data horizontally for left and right viewpoints, and so do not need the to keep track of which vertical row a source pixel moves to since horizontal movement is by definition within the same row.
- FIG. 88 shows a disparity map showing the areas where the difference between the left and right translation maps is the largest. This shows that objects closest to the viewer have pixels that are shifted the most between the two UV (or U) maps shown in Figs. 85A-B (or 86A-B).
- FIG. 89 shows a left eye rendering of the source image of FIG. 82.
- FIG. 90 shows a right eye rendering of the source image of FIG. 82.
- FIG. 91 shows an anaglyph of the images of FIG. 89 and FIG. 90 for use with Red/Blue glasses.
- FIG. 92 shows an image that has been masked and is in the process of depth enhancement for the various layers, including the actress layer, door layer, background layer (showing missing background information that may be filled in through generation of missing information - see Figs. 34, 73 and 76 for example).
- the empty portion of the background behind the actress in FIG. 92 can be filled with generated image data, (see the outline of the actress's head on the background wall).
- a compositing program for example may be utilized as opposed to re-rendering or ray- tracing all images in a scene for real-time editing. For example, if the hair mask of the actress in FIG.
- any pixels uncovered by the new mask that are obtained from the background and are nearly instantaneous available to view (as opposed to standard re-rendering or ray-tracing that can take hours of processing power to re-render all of the images in a scene when anything in a scene is edited).
- This may include obtaining generated data for any layer including the background for use in artifact free 3D image generation.
- FIG. 93 shows a UV map overlaid onto an alpha mask associated with the actress shown in FIG. 92 which sets the translation offsets in the resulting left and right UV maps based on the depth settings of the various pixels in the alpha mask.
- This UV layer may be utilized with other UV layers to provide a quality assurance workgroup (or other workgroup) with the ability to real-time edit the 3D images, for example to correct artifacts, or correct masking errors without re-rendering an entire image. Iterative workflows however may require sending the frame back to a third-world country for rework of the masks, which are then sent back to a different workgroup for example in the United States to re -render the image, which then viewed again by the quality assurance workgroup.
- This type of iterative workflow is eliminated or minor artifacts altogether since the quality assurance workgroup can simply reshape an alpha mask and regenerate the pixel offsets from the original source image to edit the 3D images in real-time and avoid involving other workgroups for example.
- Setting the depth of the actress as per FIGS. 42- 70 for example or in any other method determines the amount of shift that the unaltered UV map undergoes to generate to UV maps, one for left-eye and one for right-eye image manipulation as per FIGS. 85A-D, (or U maps in FIGS. 86A-D).
- the maps may be supplied for each layer along with an alpha mask for example to any compositing program, wherein changes to a mask for example allows the compositing program to simply obtain pixels from other layers to "add up" an image in real-time. This may include using generated image data for any layer (or gap fill data if no generated data exists for a deeper layer).
- a set of layers with masks are combined in a compositing program to form an output image by arbitrating or otherwise determining which layers and corresponding images to lay on top of one another to form an output image. Any method of combining a source image pixel to form an output pixel using a pair of horizontal translation maps without re-rendering or ray-tracing again after adding depth is in keeping with the spirit of the invention.
- FIG. 94 shows a workspace generated for a second depth enhancement program, based on the various layers shown in FIG. 92, i.e., left and right UV translation maps for each of the alphas wherein the workspace allows for quality assurance personnel (or other workgroups) to adjust masks and hence alter the 3D image pair (or anaglyph) in real-time without re -rendering or ray-tracking and/or without iteratively sending fixes to any other workgroup.
- One or more embodiments of the invention may loop through a source file for the number of layers and create script that generates the workspace as shown in Fig. 94.
- the rendering group may read in the mask files programmatically and generate script code that includes generation of a source icon, alpha copy icons for each layer, left and right UV maps for each layer based on the rendering groups rendered output, and other icons to combine the various layers into left and right viewpoint images.
- script code that includes generation of a source icon, alpha copy icons for each layer, left and right UV maps for each layer based on the rendering groups rendered output, and other icons to combine the various layers into left and right viewpoint images.
- Any method for generation of a graphical user interface for a worker to enable real-time editing of 3D images including a method to create a source icon for each frame, that connects to an alpha mask icon for each layer and generates translation maps for left and right viewpoints that connect to one another and loops for each layer until combining with an output viewpoint for 3D viewing is in keeping with the spirit of the invention.
- any other method that enables real-time editing of images without re-rendering through use of a pair of translation maps is in keeping with the spirit of the invention even if the translation maps are not viewable or not shown to the user.
- FIG. 95 shows a workflow for iterative corrective workflow.
- a mask workgroup generates masks for objects, such as for example, human recognizable objects or any other shapes in an image sequence at 9501. This may include generation of groups of sub-masks and the generation of layers that define different depth regions. This step is generally performed by unskilled and/or low wage labor, generally in a country with very low labor costs.
- the masked objects are viewed by higher skilled employees, generally artists, who apply depth and/or color to the masked regions in a scene at 9502. The artists are generally located in an industrialized country with higher labor costs.
- Another workgroup generally a quality assurance group then views the resulting images at 9503 and determines if there are any artifacts or errors that need fixing based on the requirements of the particular project. If so, the masks with errors or locations in the image where errors are found are sent back to the masking workgroup for rework, i.e., from 9504 to 9501. Once there are no more errors, the process completes at 9505.
- FIG. 96 shows an embodiment of the workflow enabled by one or more embodiments of the system in that each workgroup can perform real-time editing of 3D images without re- rendering for example to alter layers/colors/masks and/or remove artifacts and otherwise correct work product from another workgroup without iterative delays associated with re -rendering/ray- tracing or sending work product back through the workflow for corrections.
- the generation of masks occurs as in Fig. 95 at 9501, depth is applied as occurs in Fig. 95 at 9502.
- the rendering group generates translation maps that accompany the rendered images to the quality assurance group at 9601.
- the quality assurance group views the work product at 9503 as in Fig. 95 and also checks for artifacts as in Fig.
- the quality assurance workgroup (or other workgroup) has translation maps, and the accompanying layers and alpha masks, they can edit 3D images in real-time or otherwise locally correct images without re- rendering at 9602, for example using commercially available compositing programs such as NUKE® as one skilled in the art will appreciate.
- the quality assurance workgroup can open a graphics program that they are familiar with (as opposed to a complex rendering program used by the artists), and adjust an alpha mask for example wherein the offsets in each left right translation map are reshaped as desired by the quality assurance workgroup and the output images are formed layer by layer (using any generated missing background information as per Figs.
- generating two output images from furthest back layer to foreground layer can be done without ray-tracing, by only overlaying pixels from each layer onto the final output images nearly instantaneously. This effectively allows for local pixel-by-pixel image manipulation by the quality assurance workgroup instead of 3D modeling and ray-tracing, etc., as utilized by the rendering workgroup. This can save multiple hours of processing time and/or delays associated with waiting for other workers to re-render a sequence of images that make up a scene.
Landscapes
- Processing Or Creating Images (AREA)
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
- Studio Devices (AREA)
Abstract
Description
Claims
Applications Claiming Priority (2)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
US13/366,899 US9031383B2 (en) | 2001-05-04 | 2012-02-06 | Motion picture project management system |
PCT/US2013/035506 WO2013120115A2 (en) | 2012-02-06 | 2013-04-05 | Motion picture project management system |
Publications (2)
Publication Number | Publication Date |
---|---|
EP2812894A2 true EP2812894A2 (en) | 2014-12-17 |
EP2812894A4 EP2812894A4 (en) | 2016-04-06 |
Family
ID=48948173
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
EP13746718.9A Withdrawn EP2812894A4 (en) | 2012-02-06 | 2013-04-05 | Motion picture project management system |
Country Status (5)
Country | Link |
---|---|
EP (1) | EP2812894A4 (en) |
CN (1) | CN104272377B (en) |
AU (1) | AU2013216732B2 (en) |
CA (1) | CA2866672A1 (en) |
WO (1) | WO2013120115A2 (en) |
Families Citing this family (11)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US10724861B2 (en) * | 2015-08-03 | 2020-07-28 | Commonwealth Scientific And Industrial Research Organisation | Monitoring systems and methods |
US10122996B2 (en) * | 2016-03-09 | 2018-11-06 | Sony Corporation | Method for 3D multiview reconstruction by feature tracking and model registration |
US10257501B2 (en) * | 2016-04-06 | 2019-04-09 | Facebook, Inc. | Efficient canvas view generation from intermediate views |
CN107330410B (en) * | 2017-07-03 | 2020-06-30 | 南京工程学院 | Anomaly detection method based on deep learning in complex environment |
CN109218750B (en) * | 2018-10-30 | 2022-01-04 | 百度在线网络技术(北京)有限公司 | Video content retrieval method, device, storage medium and terminal equipment |
CN111526422B (en) * | 2019-02-01 | 2021-08-27 | 网宿科技股份有限公司 | Method, system and equipment for fitting target object in video frame |
CN110060213B (en) * | 2019-04-09 | 2021-06-15 | Oppo广东移动通信有限公司 | Image processing method, image processing device, storage medium and electronic equipment |
CN112183629B (en) * | 2020-09-28 | 2024-05-28 | 海尔优家智能科技(北京)有限公司 | Image identification method and device, storage medium and electronic equipment |
TWI783718B (en) * | 2021-10-07 | 2022-11-11 | 瑞昱半導體股份有限公司 | Display control integrated circuit applicable to performing real-time video content text detection and speech automatic generation in display device |
CN115188091B (en) * | 2022-07-13 | 2023-10-13 | 国网江苏省电力有限公司泰州供电分公司 | Unmanned aerial vehicle gridding inspection system and method integrating power transmission and transformation equipment |
CN115131409B (en) * | 2022-08-26 | 2023-01-24 | 深圳深知未来智能有限公司 | Intimacy matrix viewpoint synthesis method, application and system based on deep learning |
Family Cites Families (19)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
DE3736790A1 (en) * | 1987-10-30 | 1989-05-11 | Broadcast Television Syst | METHOD FOR AUTOMATICALLY CORRECTING IMAGE ERRORS IN FILM SCANNING |
US5328073A (en) * | 1992-06-24 | 1994-07-12 | Eastman Kodak Company | Film registration and ironing gate assembly |
US5835163A (en) * | 1995-12-21 | 1998-11-10 | Siemens Corporate Research, Inc. | Apparatus for detecting a cut in a video |
US5841512A (en) * | 1996-02-27 | 1998-11-24 | Goodhill; Dean Kenneth | Methods of previewing and editing motion pictures |
US5778108A (en) * | 1996-06-07 | 1998-07-07 | Electronic Data Systems Corporation | Method and system for detecting transitional markers such as uniform fields in a video signal |
US5959697A (en) * | 1996-06-07 | 1999-09-28 | Electronic Data Systems Corporation | Method and system for detecting dissolve transitions in a video signal |
US5920360A (en) * | 1996-06-07 | 1999-07-06 | Electronic Data Systems Corporation | Method and system for detecting fade transitions in a video signal |
US5767923A (en) * | 1996-06-07 | 1998-06-16 | Electronic Data Systems Corporation | Method and system for detecting cuts in a video signal |
US6067125A (en) * | 1997-05-15 | 2000-05-23 | Minerva Systems | Structure and method for film grain noise reduction |
US6031564A (en) * | 1997-07-07 | 2000-02-29 | Reveo, Inc. | Method and apparatus for monoscopic to stereoscopic image conversion |
US8401336B2 (en) * | 2001-05-04 | 2013-03-19 | Legend3D, Inc. | System and method for rapid image sequence depth enhancement with augmented computer-generated elements |
EP1483909B1 (en) * | 2002-03-13 | 2010-04-28 | Imax Corporation | Systems and methods for digitally re-mastering or otherwise modifying motion pictures or other image sequences data |
US8156532B2 (en) * | 2002-07-15 | 2012-04-10 | Sony Corporation | Video program creation system, table providing device, terminal device, terminal processing method, program, recording medium |
EP1420407B1 (en) * | 2002-11-15 | 2012-01-04 | Sony Corporation | Method and apparatus for controlling editing image display |
US9047915B2 (en) * | 2004-04-09 | 2015-06-02 | Sony Corporation | Asset revision management in media production |
US7110605B2 (en) * | 2005-02-04 | 2006-09-19 | Dts Az Research, Llc | Digital intermediate (DI) processing and distribution with scalable compression in the post-production of motion pictures |
US20090196570A1 (en) * | 2006-01-05 | 2009-08-06 | Eyesopt Corporation | System and methods for online collaborative video creation |
US8443284B2 (en) * | 2007-07-19 | 2013-05-14 | Apple Inc. | Script-integrated storyboards |
US8225228B2 (en) * | 2008-07-10 | 2012-07-17 | Apple Inc. | Collaborative media production |
-
2013
- 2013-04-05 CA CA2866672A patent/CA2866672A1/en not_active Abandoned
- 2013-04-05 EP EP13746718.9A patent/EP2812894A4/en not_active Withdrawn
- 2013-04-05 AU AU2013216732A patent/AU2013216732B2/en not_active Ceased
- 2013-04-05 WO PCT/US2013/035506 patent/WO2013120115A2/en active Application Filing
- 2013-04-05 CN CN201380018690.7A patent/CN104272377B/en not_active Expired - Fee Related
Also Published As
Publication number | Publication date |
---|---|
EP2812894A4 (en) | 2016-04-06 |
CN104272377A (en) | 2015-01-07 |
CA2866672A1 (en) | 2013-08-15 |
AU2013216732B2 (en) | 2014-10-02 |
CN104272377B (en) | 2017-03-01 |
WO2013120115A2 (en) | 2013-08-15 |
AU2013216732A1 (en) | 2014-09-25 |
WO2013120115A3 (en) | 2013-10-24 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
US9595296B2 (en) | Multi-stage production pipeline system | |
US9615082B2 (en) | Image sequence enhancement and motion picture project management system and method | |
US8385684B2 (en) | System and method for minimal iteration workflow for image sequence depth enhancement | |
US9031383B2 (en) | Motion picture project management system | |
AU2013216732B2 (en) | Motion picture project management system | |
US8897596B1 (en) | System and method for rapid image sequence depth enhancement with translucent elements | |
US8078006B1 (en) | Minimal artifact image sequence depth enhancement system and method | |
US8396328B2 (en) | Minimal artifact image sequence depth enhancement system and method | |
CA2446150C (en) | Image sequence enhancement system and method | |
AU2015213286B2 (en) | System and method for minimal iteration workflow for image sequence depth enhancement |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
PUAI | Public reference made under article 153(3) epc to a published international application that has entered the european phase |
Free format text: ORIGINAL CODE: 0009012 |
|
17P | Request for examination filed |
Effective date: 20140905 |
|
AK | Designated contracting states |
Kind code of ref document: A2 Designated state(s): AL AT BE BG CH CY CZ DE DK EE ES FI FR GB GR HR HU IE IS IT LI LT LU LV MC MK MT NL NO PL PT RO RS SE SI SK SM TR |
|
AX | Request for extension of the european patent |
Extension state: BA ME |
|
A4 | Supplementary search report drawn up and despatched |
Effective date: 20160309 |
|
RIC1 | Information provided on ipc code assigned before grant |
Ipc: G09G 5/18 20060101ALI20160303BHEP Ipc: H04N 21/60 20110101ALI20160303BHEP Ipc: G06F 17/30 20060101ALI20160303BHEP Ipc: G06Q 10/10 20120101AFI20160303BHEP Ipc: G11B 27/34 20060101ALI20160303BHEP |
|
DAX | Request for extension of the european patent (deleted) | ||
STAA | Information on the status of an ep patent application or granted ep patent |
Free format text: STATUS: THE APPLICATION IS DEEMED TO BE WITHDRAWN |
|
18D | Application deemed to be withdrawn |
Effective date: 20171103 |