KR102589194B1 - Motion cover method, apparatus, and system for linking animation data between different characters in a compatible manner - Google Patents

Abstract

A motion cover system for linking animation data in a compatible manner may include a processor and memory. The processor: receives a first image corresponding to a first character from a user terminal and receives a second image corresponding to a second character; extracts skeletal model data from a source character; divides the extracted skeletal model data into a parent joint that is not affected by the movement of a child joint and a child joint that is affected by the movement of the parent joint based on the inverse kinematics (IK) function, forward kinematics (FK) function, and preset data for the human skeleton; determines a skeleton based on the shape of the first image, determines a parent joint on the determined skeleton, and then matches it with the parent joint of the source character; determines the movement of the child joint of the first character based on the movement of the child joint according to the parent joint of the source character; observes the movement of the source character in real time around the corresponding parent joint, and applies the movement of the parent joint of the source character to the parent joint of the first character; determines a skeleton based on the shape of the second image, determines a parent joint on the determined skeleton, and matches it with the parent joint of the first character; determines the movement of the child joint of the second character based on the movement of the child joint according to the parent joint of the first character; and may apply the movement of the parent joint of the first character to the parent joint of the second character. According to embodiments of the present invention, it is possible to generate motions of multiple characters and create content by applying motion data to characters of various body types.

Description

Motion cover method, apparatus and system for compatible linking animation data between different characters {MOTION COVER METHOD, APPARATUS, AND SYSTEM FOR LINKING ANIMATION DATA BETWEEN DIFFERENT CHARACTERS IN A COMPATIBLE MANNER}

This document relates to electronic devices and systems, and specifically to a motion cover method, device, and system for compatiblely connecting animation data between different characters.

Character motions in traditional 3D computer animation were created using animators' keyframing techniques, and keyframing techniques are still one of the most representative methods of producing character animation in the film and animation industry. To produce such keyframing animations, animators mostly work with elaborately designed character rigs.

Character rigs generally refer to intuitive and familiar controllers created to allow animators to more conveniently manipulate a character's joint motion and natural body deformation for animation production. Animators use these controllers to control the parameters of the character's joints and deformers, making them useful for creating new keyframing animations or modifying already created animations.

Since the keyframing method relies heavily on the animator's capabilities and has the disadvantage of being time-consuming and expensive, motion capture and retargeting technologies that can obtain realistic motion more easily and quickly are receiving attention.

Thanks to many inventions and technological developments over the years, motion capture and retargeting technologies are now very important in real-world game, film, and animation production, along with traditional keyframing techniques.

Motion retargeting increases the reusability of motion capture data by applying a character's motion to other environments or characters. Because of its importance, it has been extensively studied in relation to computer graphics. Many existing methods focused on motion retargeting for characters of various sizes, topologies, and various terrains. Most of these methods aim to make the input motion natural by changing the input motion for changed characters or terrain and to relatively reduce the complexity of the input motion from an interaction perspective.

As the input motion contains more complex semantics, such as complex multiperson interactions, the motion retargeting problem can become more difficult because the resulting motion must not only look realistic but also maintain the meaning of the motion.

Republic of Korea Patent Publication No. 10-2020-0005894 Republic of Korea Patent Publication No. 10-2012-0020137 Republic of Korea Patent No. 10-2364728 Republic of Korea Patent Publication No. 10-2008-0050283

You can create various animations through a source character (e.g., a mannequin), copy the movements, apply them to the desired character, and then delete the source character (e.g., a mannequin). However, this method may increase the amount of data that needs to be stored in the system because the movements must be copied for each character to which it is applied. Additionally, even if a motion is not used by a specific character, it may be used by another character, so there is a limitation that it is difficult to selectively delete the data.

In addition, the existing retargeting method calculates the movement of the source character (e.g., a mannequin) based on coordinates or a mesh, extracts the data, stores it separately, and uses a method to use it on the target character to which the source character's movement is applied. You can. In this case, a data file may be required for the operation of the target character, but if there are multiple target characters, the data of the source character (e.g., mannequin) must be individually copied for each character, which is cumbersome and slow.

The motion cover method, device, and system for interchangeably linking animation data between different characters according to the present document enable compatibility between target characters, so that the target character can be used even in situations where data of the source character (e.g., mannequin) is not possessed. Data between characters can be shared in real time and made compatible.

A motion cover system that interchangeably connects animation data may include a processor and memory.

The processor receives a first image corresponding to the first character from the user terminal, receives a second image corresponding to the second character, extracts skeletal model data from the source character, and converts the extracted skeletal model data into IK (Inverse IK). Based on the Kinematics (FK) function, FK (Forward Kinematics) function, and preset data on the human skeleton, it is divided into a parent joint that is not affected by the movement of the child joint and a child joint that is affected by the movement of the parent joint. 1 Determine a skeleton based on the shape of the image, determine a parent joint on the determined skeleton, match it with the parent joint of the source character, and determine the child of the first character based on the movement of the child joint according to the parent joint of the source character. Determine the movement of the joint, observe the movement of the source character in real time around the corresponding parent joint, apply the movement of the parent joint of the source character to the parent joint of the first character, and Determining a skeleton based on the shape, determining a parent joint on the determined skeleton, matching it with the parent joint of the first character, and establishing a child joint of the second character based on the movement of the child joint according to the parent joint of the first character. The movement of can be determined, and the movement of the parent joint of the first character can be applied to the parent joint of the second character.

According to embodiments of the present invention, complex interaction movements can be motion retargeted to characters of different body types, and this can be applied to movies and 3D games where movements must be created to suit various characters, greatly reducing the time for character movement creation. You can.

According to embodiments of the present invention, it is possible to generate motions of multiple characters and produce content by applying motion data to characters of various body types.

According to embodiments of the present invention, it can be provided as motion creation software for animators in movies and 3D games, and can be used as a base technology for creating content using the user's real-time virtual avatar character.

Figure 1 is a diagram showing a conventional animation production pipeline.
Figures 2 and 3 are diagrams showing a conventional user's skeletal model.
Figure 4 is a block diagram showing the configuration of a system for connecting animation data between different characters interchangeably according to an embodiment.
Figure 5 is a flowchart showing a method of connecting animation data between different characters to be compatible, according to an embodiment.
Figure 6 is a flowchart showing a method of connecting animation data between different characters in a compatible manner according to an embodiment.

Figure 1 is a diagram showing a conventional animation production pipeline.

As shown in Figure 1, whenever a storyboard is published, many workers recreate similar or identical actions to fit the scene due to various variables such as the character's position, interactive objects, and camera angle. So, an animation is being produced. In other words, although it would be efficient to produce only the necessary movements, it is operated as a labor-intensive business that requires many production stages and a large number of manpower to complete the animation.

Accordingly, the movements of animation works with similar concepts are collected, the movement patterns are analyzed, frequently used movements and movements for each part are patterned, and these movements can be mixed and recycled according to the situation for each part, and the animation movement There is an increasing demand to reduce inefficient repetitive work processing by allowing partial modification.

In particular, series animations for TV often have the same characters appear in the same background even though there are many episodes, with only the story development being different. Therefore, productivity can be greatly increased by databaseizing only the character movements, backgrounds, and camera values that appear in the entire series and reusing them for each part.

Meanwhile, in the existing animation method, when the storyboard is completed, each scene is distributed to several animation workers and the completed scenes are reassembled to create the final video. In the process, workers often use the wrong characters or backgrounds, and each scene is Another problem is that it takes a lot of time to set up.

In order to solve this problem, there is a need for technology to automatically load the characters recorded for each scene on the storyboard, the background appropriate for the scene, camera settings, and lighting values appropriate for day and night without manual work. In other words, there is an increasing demand to save time by setting pre-set values or creating in advance a set of basic hand movements, which are partial movements, as well as walking and running movements that are essential for animation characters.

In addition, when creating a storyboard, there is an increasing demand to create a storyboard using already created actions and to store all newly added actions in the database and use them in the next episode. Ultimately, when the storyboard is created after many rounds, animation is created using the animation in the database, and research is required on technology that allows animation to be produced with only partial detail modifications.

Figures 2 and 3 are diagrams showing a conventional user's skeletal model.

Figure 2 shows an example skeletal mapping of a user that can be generated from a capture device. In this embodiment, various joints and bones, namely each hand 202, each forearm 204, each elbow 206, each bicep 208, each shoulder 210, each hip 212, each thigh 214, each knee 216, each forelimb 218, each foot 220, head 222, middle spine 224, upper portion of spine ( 226) and lower part 228 and waist 430 are identified. When more points are tracked, additional features can be identified, such as the bones and joints of the fingers or toes, or individual features of the face such as the nose and eyes. It is understood that one or more of the points shown in Figure 2 may be omitted and/or others may be added. Furthermore, skeletal mapping is an example of a computer model of a user; other computer models may be considered.

Each of the points in Figure 2 can be described in three-dimensional Cartesian space by x, y and z coordinates in a frame of reference for the capture device (camera space).

In one embodiment, the upstream joint is the next adjacent joint closer to the torso. Thus, the joint upstream of the wrist is the elbow, the joint upstream of the elbow is the shoulder, the joint upstream of the ankle is the knee, and the joint upstream of the knee is the hip.

The prior art can provide target recognition, motion analysis and tracking systems with the ability to recognize and interpret relatively complex gestures such as dance steps and the like.

In one embodiment, specific motions corresponding to dance steps are evaluated based on user input data regarding the motion of at least some of the skeletal points shown in FIG. 2. The relative motion of these points is evaluated over a fixed period of time (e.g., 8 beats of music) and compared to stored prototype data representing a catalog of specific prototype moves or steps. A determination may then be made as to which particular dance step or move is being performed by the user. Individual prototype steps or moves are referred to herein as motion classes. Additionally, correlation of the user's measured movements and energy to prototype data can be used to provide a score indicative of the quality of the user's interpretation of the specific dance move being performed.

Points on the human torso rarely exhibit strong independent motion. Therefore, the torso can be considered as a vertically extending rigid body. However, it can be seen that due to strong noise patterns in the depth sensing system, individual torso points, especially the shoulders and hips, may exhibit unrealistic motion, which would preferably be constrained rather than propagated by the relative representation. . As a result, the body can be regarded as a rigid body where all of its points contribute to the estimate of its position. These estimates can then be used to represent the remainder of the human skeleton in a relative manner.

Figure 3 is a diagram showing a conventional user's skeletal model.

Determine inverse kinematics for retargeting the source motion to the target character. All characters can have the same joint hierarchy, with a certain number (e.g. 20) of joints, including hands, of different sizes and skin shapes.

Kinect recognizes the human skeleton by finding the positions of a certain number (e.g. 20) of joints using NUI (Natural User Interface) API. In Kinect to which the present invention is applied, the joints are recognized through NUI. A person can be recognized based on their state of movement. NUI refers to a method of communicating directly through natural human body movements rather than mechanical devices such as a mouse or keyboard.

Each joint may include, for example, HEAD, SHOULDER_CENTER, SPINE, HIP_CENTER, and the SHOULDER, ELBOW, WRIST, HAND, HIP, KNEE, ANKLE, and FOOT of the arm. Since there are 4 joints in the arm and 4 joints in the leg, 2 each, Kinect can track the positions and information of a total of 20 joints.

The system according to this document can recognize a person's posture using a skeleton and separate human areas using a depth image. The positions of joints can be indicated by small circles. After recognizing the motion, the motion can be stored using the skeleton position array in the skeleton data structure.

Based on Kinect's skeletal information, it is possible to generate transformation matrix information based on the relative positions of parent and child joints for calculating the direction of the joint from the position of the joint for motion processing of the object. To obtain the direction of a joint from the position of a given joint, a transformation matrix based on the relative positions of the parent and child joints may be needed. A parent joint refers to a joint that is not affected by the position or direction of the child joint. A child joint refers to a joint that is affected by the position or direction of the parent joint.

The retargeted motion must follow the source motion while maintaining the spatial relationship between interacting characters. Since the shape of the target character is different from the shape of the source character, adjustment may be necessary.

Figure 4 is a block diagram showing the configuration of a system for connecting animation data between different characters interchangeably according to an embodiment.

According to FIG. 4, the system 400 may include a processor 410 and a memory 420, and some of the depicted components may be omitted or replaced. System 400 according to one embodiment may be a server or a terminal. According to one embodiment, the processor 410 is a component capable of performing operations or data processing related to control and/or communication of each component of the system 400 and may be composed of one or more processors. The memory 420 may store information related to the above-described method or store a program implementing the above-described method. Memory 420 may be volatile memory or non-volatile memory. The memory 420 can store various file data, and the stored file data can be updated according to the operation of the processor 410.

According to one embodiment, the processor 410 can execute a program and control the system 400. The code of the program executed by the processor 410 may be stored in the memory 420. Operations of the processor 120 may be performed by loading instructions stored in the memory 420. The system 400 is connected to an external device (eg, a personal computer or a network) through an input/output device (not shown) and can exchange data.

According to one embodiment, there will be no limitation to the calculation and data processing functions that the processor 410 can implement on the system 400, but hereinafter, the function of compatiblely connecting animation data between different characters will be described. will be.

Figure 5 is a flowchart showing a method of connecting animation data between different characters to be compatible, according to an embodiment.

Although the process steps, method steps, algorithms, etc. are depicted in a sequential order in the flow chart of FIG. 5, such processes, methods, and algorithms may be configured to operate in any suitable order. In other words, the steps of the processes, methods and algorithms described in various embodiments of the invention do not need to be performed in the order described herein. Additionally, although some steps are described as being performed asynchronously, in other embodiments, some such steps may be performed concurrently. Additionally, illustration of a process by depiction in the drawings is not intended to imply that the illustrated process excludes other variations and modifications thereto, and that any of the illustrated process or steps thereof may be incorporated into any of the various embodiments of the invention. It does not imply that more than one is required, nor does it imply that the illustrated process is preferred.

In operation 510, a processor (eg, processor 410 of FIG. 4) may receive a first image corresponding to the first character and a second image corresponding to the second character. The processor 410 may receive a first image corresponding to the first character and a second image corresponding to the second character from the user terminal. The first character and the second character may mean different types of characters.

At operation 520, processor 410 may extract skeletal model data on the source character. For example, the source character may mean a mannequin that is the target of motion extraction. In one embodiment, a source character may be used to extract motion.

In operation 530, the processor 410 may distinguish between a parent joint and a child joint based on the extracted skeletal model data. The processor 410 converts the extracted skeleton model data into a parent joint that is not affected by the movement of the child joint and the parent based on the IK (Inverse Kinematics) function, the FK (Forward Kinematics) function, and preset data for the human skeleton. It can be divided into child joints that are affected by the movement of the joint.

The IK (Inverse Kinematics) function may refer to the function of fixing a bone to a character and moving the controller at the end of the bone. Inverse Kinematics (IK) functions can be used on joints related to the foot (e.g. ankles) to give a fixed feel, for example, when a character has the soles of their feet on the ground. The FK (Forward Kinematics) function may refer to a function that provides an animation effect by moving the rotation value of a bone. The Forward Kinematics (FK) function can be used, for example, to give the feeling of being unstuck because the feet are not fixed in a situation where a character is floating in the air or moving through water.

In operation 540, the processor 410 may determine the skeleton based on the shape of the first image and correspond to the parent joint. The processor 410 may refer to a parent joint as a joint that is not affected by the movements of other joints, unlike child joints. The processor 410 may first determine the position of the parent joint and then determine the relative positions of the surrounding child joints based on this position. The relative positions of the parent joint and child joint in the first character may not be fixed. The relative positions of parent joints and child joints may vary depending on the joint positions on the source character. For example, in the case of a child joint located at an angle of 45 degrees to the left of the top of the parent joint, this position is not fixed, and the placement may vary depending on the relative joint position of the source character. That is, if a child joint located at an angle of 45 degrees to the left with respect to the top of the parent joint on the source character moves to the right, the child joint on the first character may also move to the right with respect to the parent joint. The number of parent joints and the number of child joints is not limited, but the location and name of the joints can be determined based on the shape of a typical person. Joints for a general human shape have been previously described in FIGS. 1 to 3.

In operation 550, the processor 410 may determine the movement of the child joint of the first character based on the movement of the child joint according to the parent joint of the source character.

At operation 560, the processor 410 may observe the movement of the source character about the parent joint in real time and apply the movement on the parent joint of the first character. Rather than storing the movement of the source character as data and applying it to the joints of the first character after calculation, the processor 410 may observe the movement of the source character in real time and apply it to the joints of the first character. When performing calculations by storing the movement of the source character as data, the movement of the source character must be observed in advance, and since it must be stored as data, a relatively large amount of storage space is required, and analysis is carried out only after observation is completed. The time to cover motion based on the character's movement may be delayed. On the other hand, the motion cover system 400, which interchangeably connects animation data according to this document, can reduce the space for storing data by observing the movement of the source character in real time and applying it to the joints of the first character, and Since the movement is transmitted in real time in a moving situation, motion cover can be performed relatively faster.

In operation 570, the processor 410 may apply operations 510 to 560 to the second character. The processor 410 may determine a skeleton based on the shape of the second image, determine a parent joint on the determined skeleton, and then correspond to the parent joint of the first character. The processor 410 determines the movement of the child joint of the second character based on the movement of the child joint according to the parent joint of the first character, and matches the movement of the parent joint of the first character to the parent joint of the second character. It can be applied to .

Figure 6 is a flowchart showing a method of connecting animation data between different characters in a compatible manner according to an embodiment.

At operation 610, a processor (e.g., processor 410 in FIG. 4) may determine the positions of the skeleton and joints corresponding to the initial pose of the source character.

In operation 620, the processor 410 may change the initial pose of the first image. The processor 410 may change the pose of the first image to match or correspond to the pose of the source character. This is basic work to increase the accuracy of motion cover and can be applied even when, for example, the source character is a human and the first character type is a quadrupedal creature. If the type of the first character is a bipedal creature, the joints can be arranged to correspond to the shape of the source character, a person. However, if the first character type is a creature that walks on four legs, it may be difficult to arrange the joints to correspond to the shape of the source character, a person. The processor 410 may transform the first image to match the initial posture with the source character even if the type of the first character is a creature that walks on four legs. The motion cover system that connects animation data interchangeably according to this document can control the initial posture to facilitate covering even when the type of the first character is not a living creature such as a person walking on two legs.

In operation 630, the processor 410 may determine the positions of the parent joint and child joint of the first character. The processor 410 may determine the positions of the parent and child joints of the first character based on the posture of the first image and the positions of the parent and child joints of the source character.

In operation 640, the processor 410 may check whether the joint position of the first character corresponds to the joint position of the source character. It is possible to check whether the joint position of the first character corresponds to the joint position of the source character using the tag corresponding to the joint of the source character. A tag may mean, for example, the name of a bone or a number corresponding to the bone.

In operation 650, the processor 410 may re-determine the positions of the parent joint and child joint of the first character based on changes in the positions of the skeleton and joints of the source character. The processor 410 may change the posture of the first image based on changes in the positions of the skeleton and joints of the source character and re-determine the positions of the parent joints and child joints of the first character. Additionally, for the second character corresponding to the second image, the positions of the parent joint and child joint of the second character can be adjusted similarly.

Additionally, based on the skeleton and joint movements of the first character, the skeleton and joints of the second character may be differently controlled and motion cover may be performed. Conversely, the skeleton and joints of the first character may be differently controlled and covered based on the skeleton and joint movements of the second character.

According to one embodiment, the processor 410 detects the motion of the source character based on at least one joint included in the source character,

Check the rotation angle for each joint when the source character moves,

When the first character makes a movement corresponding to the source character, the rotation angle is checked for each joint,

If the rotation angle of the parent joint of the source character and the rotation angle of the parent joint of the first character differ by more than a preset first level, the position of the corresponding parent joint is recorded,

If the rotation angle of the child joint of the source character and the rotation angle of the child joint of the first character differ by more than a second level that is relatively higher than the first level, the position of the corresponding child joint may be recorded. .

Here, the first level and second level are not fixed values and may vary depending on settings. If the rotation angle of the parent joint of the source character is different from the rotation angle of the parent joint of the first character, it may have a greater influence on the motion cover than the child joint. For this reason, the parent joint can be determined to need adjustment by recording its position (e.g. first level) even if the rotation angle difference is relatively small. On the other hand, child joints have relatively less influence on the motion cover than their parent joints, so only when the rotation angle difference is relatively larger (e.g. at the second level) can the position be recorded and an adjustment determined as necessary.

The processor 410 may assign a first weight to the number of parent joints exceeding the first level, and may assign a relatively lower weight than the first weight to the number of child joints exceeding the second level. there is. When it comes to motion cover, parent joints can have a relatively greater influence than child joints. And in order to decide whether to perform the motion cover again as a whole, a comprehensive score is needed. When determining this overall score, the score can be determined based on the number of joints that require adjustment. At this time, since the parent joint can have a relatively large influence on the motion cover, a larger weight can be applied. The first weight is not a fixed value and may vary depending on settings. However, a relatively lower weight is applied to the child joint than the first weight, which can reduce its influence on the overall score compared to the parent joint.

The processor 410 assigns a relatively higher weight as the position of the parent joint exceeding the first level is closer to the head or center of the first character, and the closer the position of the parent joint exceeding the first level is to the head or center of the first character. The further away from , the relatively lower weight can be assigned.

In general, people can check the field of view based on the character's head or center. Under this assumption, the closer the position of the parent joint is to the head or center of the first character, the difference in rotation angle may have a relatively large impact on the user watching. For this reason, the closer the position of the parent joint is to the head or center of the first character, the higher the weight can be assigned to increase the influence on the overall score.

The processor 410 calculates the number of parent joints exceeding the first level to which the first weight is applied, the number of child joints exceeding the second level to which the second weight is applied, and the head or center of the first character. The evaluation value is determined by considering the weight based on , and if the evaluation value exceeds a preset level, it is determined that the motion cover is not properly achieved, and the positions of the parent joint and child joint of the first character can be adjusted again. . Here, the evaluation value may refer to the overall score mentioned above. The processor 410 may determine whether motion cover has been successfully achieved based on the evaluation value. For example, the processor 410 may determine that the difference in motion of the source character is large based on the evaluation value exceeding a specified level (e.g., 10 points) and that the motion cover is not well achieved.

According to one embodiment, the processor 410 detects the motion of the source character based on at least one joint included in the source character,

Check the rotation angle for each joint when the source character moves,

When the first character makes a movement corresponding to the source character, the rotation angle is checked for each joint,

If the rotation angle of the parent joint of the source character and the rotation angle of the parent joint of the first character differ by more than a preset first level, the position of the corresponding parent joint is recorded,

If the rotation angle of the child joint of the source character and the rotation angle of the child joint of the first character differ by more than a second level that is relatively higher than the first level, the position of the corresponding child joint is recorded, and

If the evaluation value exceeds a preset level, it is determined that the motion cover is not well achieved, and when the positions of the parent joint and child joint of the first character are adjusted again, the recorded parent joint and child joint of the first character Using the location, readjust the joints at that location,

After performing readjustment, the rotation angle can be measured only for the joints recorded among the joints of the first character and compared with the rotation angle of the joints at the corresponding positions among the joints of the source character to verify whether the motion cover has been properly achieved. .

According to one embodiment, the processor 410 applies an IK (Inverse Kinematics) function when the position of the parent joint exceeding the first level is farther than the specified level based on the head or center of the first character. thereby relatively reducing the rotation angle of the parent joint that exceeds the first level,

If the position of the parent joint exceeding the first level is located close to the head or center of the first character within a specified level, the originally applied function of IK (Inverse Kinematics) or FK (Forward Kinematics) may be maintained.

The IK (Inverse Kinematics) function refers to the function of fixing a bone to a character and moving the controller at the end of the bone, and the FK (Forward Kinematics) function may refer to the function of applying an animation effect by moving the rotation value of the bone. .

The embodiments of this document disclosed in the specification and drawings are merely presented as specific examples to easily explain the technical content according to the embodiments of this document and to aid understanding of the embodiments of this document, and limit the scope of the embodiments of this document. That's not what I want to do. Therefore, the scope of an embodiment of this document should be interpreted as including all changes or modified forms derived based on the technical idea of an embodiment of this document in addition to the embodiments disclosed herein.

Claims (3)

In a motion cover system that connects animation data interchangeably,
processor; and
contains memory,
The processor is
Receiving a first image corresponding to a first character from a user terminal, receiving a second image corresponding to a second character,
Extract skeletal model data from the source character,
The extracted skeletal model data is based on the IK (Inverse Kinematics) function, FK (Forward Kinematics) function, and preset data for the human skeleton to determine the parent joint that is not affected by the movement of the child joint and the movement of the parent joint. Divided into child joints that receive
Determining a skeleton based on the shape of the first image, determining parent joints on the determined skeleton, and then matching them with parent joints of the source character,
Determining the movement of the child joint of the first character based on the movement of the child joint according to the parent joint of the source character,
Observe the movement of the source character in real time around the corresponding parent joint, and apply the movement of the parent joint of the source character to the parent joint of the first character,
Determining a skeleton based on the shape of the second image, determining parent joints on the determined skeleton, and then matching them with parent joints of the first character,
Determining the movement of the child joint of the second character based on the movement of the child joint according to the parent joint of the first character,
Applying the movement of the parent joint of the first character to the parent joint of the second character,
Detecting motion of the source character based on at least one joint included in the source character,
Check the rotation angle for each joint when the source character moves,
When the first character makes a movement corresponding to the source character, the rotation angle is checked for each joint,
If the rotation angle of the parent joint of the source character and the rotation angle of the parent joint of the first character differ by more than a preset first level, the position of the corresponding parent joint is recorded,
If the rotation angle of the child joint of the source character and the rotation angle of the child joint of the first character differ by more than a second level that is relatively higher than the first level, the position of the corresponding child joint is recorded,
A first weight is assigned to the number of parent joints exceeding the first level,
A relatively lower weight than the first weight is given to the number of child joints exceeding the second level,
As the position of the parent joint exceeding the first level is closer to the head or center of the first character, a relatively higher weight is given, and as the position of the parent joint exceeding the first level is farther from the head or center, a relatively higher weight is given. A low weight is given to
The number of parent joints exceeding the first level to which the first weight is applied, the number of child joints exceeding the second level to which a relatively lower weight than the first weight is applied, and the head or center of the first character Determine the evaluation value by considering the weight based on
If the evaluation value exceeds a preset level, it is determined that the motion cover is not properly achieved, and the positions of the parent joint and child joint of the first character are adjusted again,
If the position of the parent joint exceeding the first level is farther than the specified level based on the head or center of the first character, the IK (Inverse Kinematics) function is applied to position the parent joint exceeding the first level. Relatively reduce the rotation angle,
A system that maintains the originally applied function of IK (Inverse Kinematics) or FK (Forward Kinematics) when the position of the parent joint exceeding the first level is located close to within a specified level based on the head or center of the first character.
According to clause 1,
The processor is
Determine the positions of the skeleton and joints corresponding to the initial posture of the source character,
Changing the posture of the first image to correspond to the initial posture of the source character,
Determining the positions of the parent joints and child joints of the first character based on the posture of the first image and the positions of the parent joints and child joints of the source character,
Check whether the joint position of the first character corresponds to the joint position of the source character using a tag corresponding to the joint of the source character, and
A system that changes the posture of the first image based on changes in the positions of the skeleton and joints of the source character and re-determines the positions of the parent joints and child joints of the first character.



delete