JP2007208651A - Content viewing apparatus - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To shorten viewing time by eliminating the display of events unnecessary for a viewer. <P>SOLUTION: This content viewing device A comprises: a content input part 1; a video/audio/data separation part 3 which separates video/audio data and meta data; a highlight segment extraction part 7 which extracts segment information of a highlight part based on the meta data and the taste data; an operation part 23; a prosperity degree measurement part 21 which measures a degree of prosperity in a program; a user taste update part 17 which updates and stores the user taste data; a video/audio/data combining part 11 which combines a highlight segment and the separated video/audio data and a display part 15. In the above configuration, matching as mentioned later is performed by the meta data received by separating video/audio from content (media such as broadcast/communication/DVD) and a retrieval formula based on viewer's tastes to display an event which satisfies conditions. <P>COPYRIGHT: (C)2007,JPO&INPIT

Description

  The present invention relates to a content viewing apparatus, and more particularly to a content viewing apparatus that automatically updates user preference data according to a content viewing style using metadata and provides viewers with more efficient time-saving viewing.

  Multi-functionalization is also progressing in content viewing apparatuses for viewing content. For example, Patent Document 1 discloses the following technique. 1) A time bar corresponding to the time axis of the video content is displayed on the screen. 2) On the time bar, an icon indicating the start position of the event scene, which is a characteristic scene in the moving image, is displayed. 3) When the viewer operates the remote control, moves the focus on the icon, and selects a desired icon, video playback can be started from the start position of the event scene corresponding to the selected icon.

JP 2001-292402 A

  However, in the device described in Patent Document 1, an icon representing an event scene is displayed from the received metadata, and a scene unfavorable for the viewer is displayed as an icon because it is in accordance with the metadata. If the metadata is used as it is, there are too many icons to be displayed, and there is a problem that it takes time and effort to skip a scene that is undesirable for the viewer.

  An object of the present invention is to prevent an unnecessary event for the viewer from being displayed and to shorten the viewing time.

  In the present invention, based on the received metadata and the viewer's preference, only events that may be the viewer's preference are displayed. Further, the present invention is characterized in that the accuracy of event display is further improved by automatically updating the search formula for the viewer's preference as needed based on the viewing style.

  Unnecessary events for the viewer are not displayed, the display screen is clean, and there is no need to skip unnecessary scenes during playback, enabling more effective time viewing.

  A content viewing apparatus according to an embodiment of the present invention will be described below with reference to the drawings. FIG. 1 is a functional block diagram showing a configuration example of a content viewing apparatus according to the present embodiment. As shown in FIG. 1, the content viewing apparatus A includes a content input unit 1 for inputting content, a video / audio / data separation unit 3 for separating video / audio data and metadata in the content, and user preferences. The segment information of the highlight part is extracted based on the user preference database 5 for storing data, the metadata separated in the video / audio / data separation unit 3, and the preference data stored in the user preference database 5. The highlight segment extraction unit 7 to be operated, the operation unit 23 operated by the user, the operation from the operation unit 23 and the highlight segment extracted by the highlight segment extraction unit 7 are measured. And a user preference based on the degree of success measured by the degree of success measurement unit 21. The user preference data update unit 17 that updates the user preferences and stores the updated user preferences in the user preference database 5, the highlight segment extracted by the highlight segment extraction unit 7, and the video / audio / data separation unit 3 The video / audio / data synthesizing unit 11 for synthesizing the data and the separated video / audio data, and the display unit 15 for displaying the output.

  In the above configuration, matching is performed as described later by using metadata obtained by separating video / audio from content (broadcasting / communication / DVD media, etc.) and a search formula based on viewer's preference. Show events that meet.

  In the present embodiment, a case where the content once recorded is reproduced is shown as an example. As an example of the metadata acquisition method, if metadata is included in the broadcast, it can be taken from the broadcast data. It is also possible to create metadata based on the actual content in the video or the voice recognition of the live broadcast, or by character recognition from the back number or telop of the video.

  FIG. 2 is a diagram illustrating a state in which highlight segments are extracted only by metadata, which is the general technique described above, and each event specified by the metadata is the content of the metadata along the time axis. It is arranged with. For example, as indicated by reference numeral 31 in FIG. 2, according to the metadata alone, in order from the previous time, the scene in which the H team in the reference numeral 51 scored three points, K generated at a certain time interval, The hit scene 53, the H attack scene 55 in the front 7 times after a certain time interval, and the G attack scene 57 in the back 9 times are extracted.

  On the other hand, FIG. 3 is a technique according to the present embodiment, and shows a state in which highlight segments are extracted based on metadata and user preference data. As is clear from comparison with FIG. Of the events specified by the metadata, the event 53 and the event 57 are excluded from the target.

  This point will be described with reference to FIGS. FIG. 4 is a diagram showing a state of processing related to highlight segment extraction centering on the highlight segment extraction unit. The relevance of the video / audio / data separation unit 3, the highlight segment extraction unit 7, and the user preference database 5 and respective data content examples are shown along the example shown in FIG. Further, a positive evaluation keyword and a negative evaluation keyword are set in advance as evaluation criteria. In the figure, H teams (+20 points), points (+10 points), attacks (+8 points), etc. as positive evaluation keywords, and K players (−10 points), G teams (−18 points) as negative evaluation keywords. And so on. The user's preference is measured while referring to this table. First, the video / audio / data separation unit 3 extracts metadata such as 3 points for H, home run for K, 7-time front H attack, and 9-time back G attack. Next, the highlight segment extraction unit 7 calculates a score, if any, based on the metadata and the evaluation criteria. The new home run is set to 0 points. Among the extracted highlight segments, those with + points, that is, H3 score (+30) and 7 times H attack (+28) are extracted. It is also possible to set a search expression for each genre (for each event) to further improve accuracy. This is effective when there are players with the same surname in different events.

  Next, how to obtain points will be described. FIG. 5 is a graph in which the volume of content, which is an index of how content rises (prosperity), is plotted on the vertical axis along the playback time t of recorded content. Since audio is recorded at the time of content reproduction even if it is not real time, it is possible to estimate the degree of excitement at a certain time by plotting the volume with the time from the start of reproduction. For example, the volume exceeds a certain threshold value in each period from time t1 to time t2 and from time t3 to time t4. It can be said that this period is a period of high prosperity. In other words, the points need to be raised during this period.

  FIG. 6 is a diagram created based on FIG. 5, where the reproduction time of the content is on the horizontal axis, and the integral value (%) of the degree of success at the reproduction time t is on the right vertical axis. As shown in FIG. 6, the integral value of the prosperity L1 of a certain content rises rapidly in the second half in this example. Note that L2 is a standard content prosperity, is a content that rises linearly with respect to time, and does not have a time zone that is particularly important. On the other hand, in L3, it can be seen that the prosperity increases until time t and does not increase much thereafter. In this case, in the content L3, the degree of success is high between time 10 and t, and a point +7 is given to this time zone from the left vertical axis. In the content L1, a high point (+10) is given in a time zone after t. In this way, points based on the degree of success are given to each content.

  FIG. 7 is a flowchart summarizing the flow of the above processing. As shown in FIG. 7, the content viewing is started (step S1), and it is determined whether or not the content has been viewed for 10 seconds or more in step S2 (note that 10 seconds is a guide and can be arbitrarily set). If it is longer than the specified time (YES), the process proceeds to step S3. If the specified time is not reached, the process proceeds to step S5, and for example, -10 points are added assuming that the user is not interested, and the process ends in step S7. In step S3, it is determined whether or not all of the content has been viewed. If all of the content has been viewed, the process proceeds to step S4, for example, +10 points are added, and the process proceeds to step S7. If not all have been viewed in step S3, the process proceeds to step S6, points are added according to the viewing content (viewing time), and the process proceeds to step S7. In this way, addition / subtraction of points is performed. Since such an evaluation process produces different results depending on circumstances, it is preferable to use, for example, the latest processing procedure for the next evaluation process. For example, a highlight segment having a point of + is extracted by the above processing (see FIG. 3).

  FIG. 8 is a diagram showing a viewing example regarding the highlight obtained by the above processing. As shown in FIG. 8, a display screen 15 and time zone information are provided. A time bar 33 capable of moving the time zone information in the time axis direction by a user operation is provided. The user can move the time bar 33 on the time axis using the highlight extraction units 41, 43, 45, 47 as a guide. Content display during the time when the time bar 33 is in focus is displayed on the display unit 15. When the corresponding time bar 33 is focused, the event information or the actual moving image is displayed with a size of, for example, about 1/4 of the screen. As a result, the content of the highlighted portion can be confirmed.

  FIG. 9 is a diagram showing an outline of processing in the user preference data update unit 17. As shown in FIG. 9, for example, when the metadata 61 of the attack of the 7th table H is added, if the highlight portion of the attack of H is viewed by +7 points in the process of FIG. It is changed to +27. Other points are not changed at this point. Accordingly, the positive evaluation and the negative evaluation are as shown in the figure, and this figure is newly stored in the user preference database 5 (update).

  Next, the entire process will be described with reference to the drawings. FIG. 10 is a diagram showing a flow of a highlight reproduction procedure, and FIG. 11 is a diagram showing a flow of a user preference database update procedure. As shown in FIG. 10, first, content is acquired in step S0. For example, the content recorded in the HDD or the content received in real time is passed to the content input unit. In step S1, video / audio / metadata is input. Next, in step S2, the metadata is separated from the video / audio data. In step S3, the separated metadata and user preference data are matched to extract the one with a high point as a highlight segment to create highlight segment metadata. In the above example, content in a time zone of 0 points or more is extracted. Next, in step S4, based on the extracted highlight segment metadata, the video / audio of the highlighted portion is highlighted and reproduced.

  In the case of receiving in real time, highlight segment metadata may be created using the latest metadata, and the portion before viewing may be highlighted and reproduced.

  As shown in FIG. 11, in step S11, an integral value of the degree of success with respect to time is calculated in the highlight segment. For example, the calculation is performed in real time from the middle of recording, and the calculation is completed when the program ends. Next, in step S12, reproduction of the highlight segment is started. In step S13, at the time of highlight segment reproduction, the point obtained from the integrated value of the prosperity when the viewing time is t1 seconds is reflected in the user preference data mainly by user operation such as skipping. If the integral value of the degree of success up to t1 seconds is high, the point reflected in the user preference increases accordingly. As shown in step S14, at the time of next viewing, highlight segment extraction is performed using the new user preference data updated in the above step.

  As described above, according to the present embodiment, the user preference data is created using the prosperity, thereby further reducing the display screen without displaying unnecessary events for the viewer. In addition, it is possible to save time and time viewing more effectively by skipping unnecessary scenes during playback.

  The present invention can be used for short-time viewing of content.

It is a functional block diagram which shows the example of 1 structure of the content viewing-and-listening apparatus by one embodiment of this invention. It is a general technique and shows a state in which highlight segments are extracted only by metadata, and each event specified by metadata is arranged along with the content of metadata along the time axis. It is the technique by this Embodiment, Comprising: It is a figure which shows a mode that a highlight segment is extracted based on metadata and user preference data. It is a figure which shows the mode of the process relevant to extraction of the highlight segment centering on a highlight segment extraction part. It is the graph which plotted on the vertical axis | shaft the volume of the content which is one parameter | index of how the content rises (prosperity) along the reproduction time t of the recorded content. FIG. 6 is a diagram created on the basis of FIG. 5, where the reproduction time of the content is on the horizontal axis, and the integral value (%) of the liveliness at the reproduction time t is on the right horizontal axis. It is the flowchart figure which put together the flow of the process from FIG. 4 to FIG. It is a figure which shows the viewing example regarding the highlight obtained from FIG. 4 to FIG. It is a figure which shows the outline | summary of the process in a user preference data update part. It is a figure which shows the flow of a highlight reproduction | regeneration procedure. It is a figure which shows the flow of the update procedure of a user preference database.

Explanation of symbols

  A ... content viewing device, 1 ... content input unit, 3 ... video / audio / data separation unit, 5 ... user preference database, 7 ... highlight segment extraction unit, 11 ... video / audio / data synthesis unit, 15 ... display unit , 17 ... User preference update unit, 21 ... Activeness measurement unit, 23 ... Operation unit.

Claims (9)

A content input unit to which content including video / audio data is input;
A user preference database that stores preference data indicating the level of user preference,
A highlight segment extraction unit that extracts a highlight segment by matching processing between the metadata of the input content and the preference data;
A highlight unit extracted by the highlight segment extraction unit, and a synthesis unit that synthesizes the video / audio data and the metadata;
A content viewing apparatus comprising: a reproduction unit that highlights and reproduces content based on an output from the synthesis unit.   The highlight segment extraction unit preferentially extracts a highlight portion corresponding to the matched metadata, the highlight portion of the content including the metadata having a high preference degree from the preference data. The content viewing / listening apparatus according to claim 1.   3. The content according to claim 1, further comprising: a separation unit that separates video / audio data and metadata in the content, and using the metadata separated by the separation unit as the metadata. Viewing device.   3. The content viewing apparatus according to claim 1, wherein the metadata is acquired based on a result of character recognition of video data or a result of voice recognition of audio data among actual contents of the content. A prosperity measuring unit that measures the prosperity indicating the degree of excitement in the content;
5. The content viewing / listening according to claim 1, further comprising: a user preference data update unit that updates user preference data based on the measurement of the popularity level in the success level measurement unit. apparatus.   The prosperity is used to reflect, in the user preference data, a point obtained from an integral value of the prosperity for each viewing time by a user operation during reproduction of the extracted highlight segment. The content viewing apparatus according to claim 5.   When the content is content received in real time, highlight segment metadata is created using the latest metadata, and the content portion before viewing is highlighted based on the highlight segment metadata. The content viewing / listening device according to claim 1, wherein the content viewing / listening device is a device. Obtaining content including metadata and video / audio data; separating the metadata from the content;
Matching the separated metadata and user preference data indicating the level of user preference;
Assigning points to the matched parts based on user preferences,
Extracting the high point portion as a highlight segment and creating highlight segment metadata,
A content reproduction method comprising: highlighting video / audio of a highlight portion based on the extracted highlight segment metadata. Calculating an integral value of the prosperity with respect to time in the highlight segment;
Reflecting the points obtained from the integrated value of the prosperity for each viewing time in the user preference data by the user's operation during highlight segment playback,
The content reproduction method according to claim 8, further comprising a step of performing highlight segment extraction using the updated new user preference data.

Images