JP2005033619A - Contents management device and contents management method - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To perform fine adjustment of the total time of the retrieval results in a desired time when a scene with which a specific keyword is associated from the content data is retrieved and extracted. <P>SOLUTION: The content data are managed by using metadata with the contents of the content data described and a processing table with a processing method described corresponding to the keyword to be used in the metadata, the processing method of content data is finely set and the total time of retrieval results are finely adjusted by using the processing table for the content data by processing the content data by using the processing table. <P>COPYRIGHT: (C)2005,JPO&NCIPI

Description

  The present invention relates to a content management apparatus and a content management method for managing metadata describing the contents of content data.

  In recent years, storing moving images as digital files has become widespread. Accordingly, it has become difficult to quickly search for a desired one from an enormous amount of video files and to manage it efficiently. In order to solve this problem, an information management method for adding index information to a moving image and a video search and video summarization method using the index information have been proposed (for example, see Patent Document 1).

  Patent Document 1 specifically shows a method of preparing keyword information or the like associated with a video in a specific time section as a file different from a moving image file and using the file for searching or the like. The format of the index information file is generally described in a widely used XML format.

  Furthermore, the MPEG-7 standard for the purpose of describing video contents has been proposed today.

An authoring system that creates metadata describing the contents of a moving image has also been proposed. In a conventional authoring program, when an operator performs an input that associates a specific video section with a specific keyword as index information, the content of the keyword is described in the metadata (for example, see Non-Patent Document 2). ).
JP-A-11-288424 Pioneer R & D Vol. 11 No. 2 (September 2001), page 27 “Development of MPEG-7 Applications”

  However, the conventional video index information management method has the following problems.

  In other words, in the conventional method, a keyword is assigned to each scene, but detailed processing content is not assigned to each scene.

  Specifically, in the conventional method, when a scene associated with a specific keyword is searched for and extracted from content data, the keyword is assigned even if the total time of the search result is set to a desired time. The method of fine-tuning is not provided because the segment of the video is discrete.

  For example, if it is necessary to provide a summary video related to “goal scene” in one minute, if three 25-second scenes with the keyword “goal scene” are searched, the total of these scenes is 1 minute 15 seconds. In this case, if all the searched scenes are provided, the specified time will be exceeded. Therefore, it is conceivable to delete one scene. In this case, the total time of the scene is 50 seconds, and the summary video that should be provided in 1 minute is shortened to 50 seconds.

  The present invention has been made in view of such a point, and an object thereof is to manage information that complements a processing method or the like that is not described in moving image index information.

  The present invention manages content data using metadata describing the content of the content data and a processing table describing a processing method corresponding to a keyword used in the metadata, and using this processing table, It was made to process.

  In the present invention, the processing method of content data can be set in detail by using the processing table for the content data. As a result, when a scene associated with a specific keyword is searched and extracted from the content data, the total time of the search result can be finely adjusted to a desired time.

  A content management apparatus according to a first aspect of the present invention includes a metadata reference unit that refers to metadata that describes content data content, and a processing table that describes processing content corresponding to a keyword used in the metadata. A table referring means to be referred to and a configuration provided are employed.

  Thereby, the content data can be managed by the metadata and the processing table. As a result, it is possible to realize search and summary generation that make use of the characteristics of the keyword words used in the metadata. Further, if the contents of the processing table are changed, the contents of the processing can be arbitrarily changed without changing the system.

  The processing table provision apparatus concerning the 2nd aspect of this invention is a processing table provision apparatus which provides the said processing table with respect to the content management apparatus as described in a 1st aspect.

  As a result, it is possible to provide a processing table describing the processing content corresponding to the keyword.

  A content data search device according to a third aspect of the present invention includes content data reference means for referring to content data composed of a plurality of scenes, and metadata reference means for referring to metadata in which the contents of the scene are described by keywords. Table reference means for referring to a processing table in which processing content corresponding to a keyword used in the metadata is described; and the content data reference means for inputting the search keyword and adding the search keyword to the scene. The scene search means for searching from the above, the processing content corresponding to the keyword assigned to the searched scene is extracted from the processing table, and the processing corresponding to the extracted processing content is performed on the searched scene And a data processing means.

  As described above, by managing the content data with the metadata and the processing table, it is possible to realize a search that makes use of the characteristics of the keyword words used in the metadata. Further, if the contents of the processing table are changed, the contents of the processing can be arbitrarily changed without changing the content data search device. This is highly convenient when processing is changed without changing the content data search apparatus itself.

  According to a fourth aspect of the present invention, in the content data search device according to the third aspect, the processing content is a process of expanding the scene, and the data processing means specifies a desired total time of the searched scene. If it is less than the time, the searched scene is expanded so that the total time becomes the specified time.

  Thereby, when the total time of the searched scene does not match the specified time, the total time of the searched scene can be adjusted to the specified time.

  According to a fifth aspect of the present invention, in the content data search device according to the fourth aspect, the process of expanding the scene is performed by searching the searched scene according to a keyword characteristic assigned to the scene. This is a process of adding a part of the scene before or after the scene.

  As a result, the scene can be expanded so that the features of the scene can be made clearer by expanding the scene by the expansion method according to the characteristics of the keyword assigned to the scene.

  According to a sixth aspect of the present invention, in the content data search device according to the fourth aspect or the fifth aspect, the scene is given importance, and the data processing means is included in the searched scene. The most important scene is stretched.

  Thereby, a scene having a lot of important contents can be expanded.

  An authoring processing apparatus according to a seventh aspect of the present invention includes content data reference means for referring to content data composed of a plurality of scenes, and table reference means for referring to a processing table that describes processing details corresponding to a predetermined keyword. , By inputting a keyword to be given to the scene, extracting the processing content corresponding to the input keyword from the processing table, and performing a description corresponding to the extracted processing content on the scene, And a data creation means for generating metadata indicating the content data.

  As a result, authoring processing using the processing table can be performed. As a result, the user keyword assignment operation can be supported, and the work efficiency can be remarkably improved. In addition, if the content of the processing table is changed, the processing content can be changed to an arbitrary one without changing the authoring processing device at all, which is convenient when changing the processing without changing the authoring processing device itself. high.

  According to an eighth aspect of the present invention, in the authoring processing apparatus according to the seventh aspect, the processing content is exclusive processing, and the data creation means applies to a scene other than the scene to which the input keyword is assigned. , A keyword having a meaning exclusive to the input keyword is assigned.

  This eliminates the need for exclusive keyword assignment, which has the advantage of reducing the amount of authoring work. Further, since it is contradictory that exclusive keywords are associated with the same scene, it is highly convenient to prohibit such contradictory operations in the authoring operation.

  According to a ninth aspect of the present invention, in the authoring processing apparatus according to the seventh aspect, the processing content is an importance level assigning process, and the data creation unit is configured to perform a predetermined process on the scene to which the input keyword is assigned. Gives importance.

  Thereby, since the importance can be automatically given, the amount of work of the user is reduced.

  According to a tenth aspect of the present invention, in the authoring processing apparatus according to the seventh aspect, the processing content is processing for assigning importance in stages, and the data creation means is the scene for assigning the input keyword. With respect to the scenes before and after, the importance is assigned stepwise with the importance of the scene to which the keyword is assigned as the vertex.

  In this way, by automatically assigning importance to the previous and next scenes that are currently being processed, it is not necessary for the user to assign importance to the previous and subsequent scenes that are currently being processed. Very convenient.

  According to an eleventh aspect of the present invention, in the authoring processing apparatus according to the tenth aspect, the processing content is processing for assigning importance with emphasis before, and the data creation means assigns the input keyword. A higher importance is assigned to a scene before the scene than a scene behind the scene to which the keyword is assigned.

  As a result, for example, with respect to a keyword having a prior focus such as “goal”, the previous scene is emphasized.

  According to a twelfth aspect of the present invention, in the authoring processing apparatus according to the tenth aspect, the process content is a process of assigning importance with emphasis on the back, and the data creation means assigns the input keyword. The higher importance is assigned to the scene behind the scene than the scene before the scene to which the keyword is assigned.

  As a result, for example, a keyword with a later emphasis such as “free kick” is emphasized in the later scene.

  An authoring processing apparatus according to a thirteenth aspect of the present invention includes a content data reference means for referencing content data composed of a plurality of scenes, a processing content corresponding to a predetermined keyword, and a processing table describing identification information for the processing content A table reference means for referring to the image, a keyword to be assigned to the scene, the identification information corresponding to the input keyword is extracted from the processing table, and the extracted identification information is described in the scene Thus, a configuration is provided that includes data creation means for generating metadata indicating the contents of the content data.

  As a result, it is possible to describe identification information corresponding to the processing content, instead of describing the processing content itself in the metadata. By describing the metadata in this way, the size of the metadata can be reduced.

  According to a fourteenth aspect of the present invention, the content data is managed using metadata describing the content data and a processing table describing a processing method corresponding to a keyword used in the metadata. This is a content management method.

  According to a fifteenth aspect of the present invention, there is provided a step of referring to content data composed of a plurality of scenes, a step of referring to metadata describing the contents of the scene with keywords, and a process corresponding to the keywords used in the metadata A step of referring to a processing table describing contents; a step of inputting a search keyword and searching the scene to which the search keyword is assigned; and the keyword assigned to the searched scene. A content data search method comprising: extracting the processing content corresponding to the processing scene from the processing table; and performing the processing corresponding to the extracted processing content on the detected scene. is there.

  According to a sixteenth aspect of the present invention, there are provided a step of referring to content data composed of a plurality of scenes, a step of referring to a processing table describing processing details corresponding to a predetermined keyword, and a keyword to be assigned to the scene. The step of inputting, the step of extracting the processing content corresponding to the input keyword from the processing table, and the description corresponding to the extracted processing content for the scene, An authoring processing method comprising: generating metadata to be displayed.

  According to a seventeenth aspect of the present invention, a computer includes a step of referring to content data including a plurality of scenes, a step of referring to metadata in which the contents of the scene are described by keywords, and a keyword used in the metadata. A step of referring to a processing table describing a corresponding processing content, a step of inputting a search keyword and searching the scene to which the search keyword is assigned from the content data, and a step assigned to the searched scene And a step of extracting the processing content corresponding to the keyword from the processing table and a step of performing processing corresponding to the extracted processing content on the detected scene. is there.

  According to an eighteenth aspect of the present invention, a step of referring to content data including a plurality of scenes, a step of referring to a processing table describing processing details corresponding to a predetermined keyword, and giving to the scene Inputting the keyword to be processed, extracting the processing content corresponding to the input keyword from the processing table, and performing description corresponding to the extracted processing content on the scene, thereby performing the content data Generating metadata indicating the contents of the program.

  Hereinafter, embodiments of the present invention will be described with reference to the accompanying drawings.

(Embodiment 1)
In Embodiment 1 of the present invention, a data management method for managing content data using metadata describing the contents of the content data and a processing table describing a processing method corresponding to a word used in the metadata, and A search method using this management method will be specifically described.

  First, the content data search system according to the first embodiment will be described with reference to FIG. FIG. 1 is a configuration diagram of the content data search system according to the first embodiment.

  The content data search system 100 uses a data management method for managing content data by using metadata describing the contents of the content data and a processing table describing a processing method corresponding to a word used in the metadata. Search for the specified part from the data and output it.

  Content data is a work of creation and copyrighted work, which may be digital or analog, but is stored in the media in an electronically searchable state. The content data format is not particularly limited, such as an html format such as a web page on the Internet, a bml format used in broadcasting, or an audio file. In the first embodiment, for convenience of explanation, it is assumed to be moving image data.

  The content data search system 100 has a configuration in which a communication processing unit 101, a summary generation unit 105, and a content data providing unit 108 are connected by a data bus 104 that is a bus for transferring data.

  The communication processing unit 101 includes a request input unit 102 and a video data transmission unit 103.

  The request input unit 102 inputs a request from the user as to what kind of data is to be searched.

  The video data sending unit 103 sends the data searched by the content data search system 100 to the user.

  Communication processing performed by the communication processing unit 101 is performed by a technique such as a LAN or the Internet.

  The first embodiment has been described with the configuration in which the communication processing unit 101 performs input / output. However, for example, when a single personal computer includes all data and operation means, communication is used. Since the input / output can be performed by connection within the system without the communication processing unit 101, the communication processing unit 101 may be omitted.

  The content data providing unit 108 includes a video data storage unit 114 that stores a plurality of video data 109 and 111, and a metadata storage unit 115 that stores metadata 110 and 112.

  In the metadata 110 and 112, a link as to which video data 109 and 111 is the metadata is described, and the video data 109 and 111 and the metadata 110 and 112 are associated with each other.

  In the first embodiment, the metadata 110 is a description regarding the video data 109, and the metadata 112 is a description regarding the video data 111.

  In addition, the content data providing unit 108 includes a table storage unit 116 that stores the processing table 113.

  The processing table 113 is a table that stores a plurality of pairs of keywords used in the metadata 110 and 112 and processing for the keywords.

  As the content data providing unit 108, an information holding medium such as a hard disk, an optical disk, or a magnetic disk can be used.

  Note that the content data providing unit 108 does not necessarily have to be a single device, and a plurality of devices may be connected via a network.

  For example, video data 1 is in the hard disk of a server overseas, video data 2 is in another server in Japan, metadata is provided by digital broadcast data broadcasting, stored in broadcast receiving equipment, and processed Even if the table is provided by another server, if they are connected via a network, it is virtually the same as having all the necessary data in one device.

  As described above, when data exists on the network in a distributed manner, a plurality of metadata corresponding to one video data and a plurality of processing tables corresponding to the metadata may exist. In such a case, the URL of the data corresponding to the inside of the video data or metadata may be clearly specified so that it can be clearly specified as one file. Alternatively, not only the file name but also a value unique to each file such as a hash value of the file may be written so that one file can be selected from a plurality of files.

  The video data 109 and 111 are digital files using an encoding format such as MPEG.

  The summary generation unit 105 is provided with a scene search unit 106 and a data management unit 107.

  The scene search unit 106 performs a scene search process according to a search request including a keyword and a specified time from the video data 109 and 111 stored in the content data providing unit 108.

  The data management unit 107 has a function as a reference unit that references (accesses) the video data 109 and 111, the metadata 110 and 112, and the processing table 113 stored in the content data providing unit 108, and further, the video data The processing according to the processing table 113 is added to 109 and 111.

  The summary generation unit 105 includes a CPU and a RAM, and operates a scene search unit 106 and a data management unit 107 which are software modules.

  Next, the metadata 110 and 112 shown in FIG. 1 will be described with reference to FIG. FIG. 2 is a diagram illustrating a specific example of the metadata 110 and 112 according to the first embodiment.

  The metadata 110 and 112 shown in FIG. 2 are part of metadata that is actually used, and are described in an XML format that conforms to the MPEG-7 standard. The metadata 110 and 112 are basically described as one file in one moving image file.

  The contents of a plurality of moving image files can be described in one metadata file, and the contents of one moving image file can be described in a plurality of metadata.

  Further, the metadata 110 and 112 can be described in another format or a unique format, even if they are not in the XML format compliant with the MPEG-7 standard.

  In addition, in the metadata 110 and 112, a Media URI indicating which moving image file is related is described. Further, in the metadata 110 and 112, the contents of the scene are described for each video section defined by the start time and end time in the moving image file, that is, for each scene.

  Referring to FIG. 2, the media URI of the corresponding moving image file is described in a portion indicated by 201.

  In addition, a definition of a scene (segment) is described in a portion indicated by 202. Also, in the portion indicated by 205, the start time and end time of the scene (which may appear as duration) are described.

  Further, in a portion indicated by 203, a keyword associated with the scene is described. When the scene keyword is described in the viewpoint as indicated by 203, that is, PointOfView, the importance indicating the importance from the viewpoint, that is, the importance value 204 is also described.

  Further, as indicated by 206, the keyword may be given as an annotation, ie, Annotation.

  In this way, metadata is configured.

  Next, FIG. 3 schematically shows a keyword assignment state in the metadata.

  A horizontal axis 309 indicates the passage of time of a certain moving image file.

  Reference numerals 301a to 301d are schematic views of a scene, where the left side of the rectangle is the start time and the length is DURATION. In this example, the plurality of scenes 301a to 301d are continuous without a gap, and the end time of one scene coincides with the start time of the next scene.

  On the vertical axis 310, keywords 311 to 314 assigned to the scenes 301a to 301c are arranged. On the right side of each keyword 311 to 314, bands 302 a to 302 f indicating the scene to which the keywords 311 to 314 are assigned are arranged. For example, since the bands 302a and 302b corresponding to the scenes 301b and 301c are arranged on the right side of the keyword “Japan representative” 311, it can be seen that the keyword 311 is assigned to the scenes 301b and 301c.

  This is a situation in which keywords are defined by PointOfView 203 or the like in the scene definition 202 in FIG. 2 in actual data. A plurality of keywords can be assigned to each scene. For example, three keywords (Osaka, goal, highlight) 312 to 314 are assigned to the scene 301d.

  In addition, it has been described above that keywords 311 to 314 can be given importance, but for example, keyword 314 is given importance 0.8 for scene 301c and importance 1. 0 is given.

  It is difficult for a user or operator who handles metadata to see the metadata itself expressed in XML, and to understand the assignment of metadata keywords and importance from a bird's-eye view. However, if the schematic diagram as shown in FIG. 3 is displayed using a display device or the like, the user and the operator can understand the keywords and importance of the metadata from a bird's-eye view, which increases convenience.

  The display as shown in FIG. 3 is also effective as a GUI (Graphical User Interface) of an authoring system that assigns keywords.

  In this display, it is easier to understand if the importance value is visually shown. For example, the darker the colors of the bands 302a to 302f indicating the assignment of keywords, the higher the degree of importance, or the arrangement of patterns and symbols, so that the degree of importance can be confirmed at a glance. Will increase. For example, in the example of FIG. 3, since the color of the band 302f is darker than the band 302e, it can be seen that the importance of the scene 301d is higher than that of the scene 301c.

  In addition, the schematic diagram of metadata shown in FIG. 3 has two characteristics: (1) scenes are continuous and do not overlap, and (2) a plurality of keywords can be assigned to each scene. However, the definition of the scene is not limited to this. FIG. 4 is a schematic diagram showing a state of giving metadata according to another definition.

  Each of 401 to 404 in FIG. 4 is one scene, and one keyword 311 to 314 of Japan representative, Osaka, goal, and highlight is associated with each scene. The scenes may overlap like the scene 401 and the scene 404. Also, there may be an interval between scenes.

  Since keywords can be freely assigned to the time interval by the scene definition method as shown in FIG. 4, in cases where there are overlapping keyword assignments, convenience may be improved both visually and in terms of work. is there.

  3 and 4 are different in the definition method of the scene, but the amount of information is equal. That is, if partitions such as 501 and 502 are provided at the start time and end time of all the scenes 301a to 301d as shown in FIG. 5 and used as scene division points, the example shown in FIG. Four scenes can be defined similarly to the three scenes 301a to 301d. If it is considered that a plurality of keywords are assigned to each scene, both FIG. 4 and FIG. 3 are exactly the same in terms of keywords assigned to a specific time interval.

  As described above, there are two ways of expressing the same metadata in terms of information amount, the method shown in FIG. 3 and the method shown in FIG. 4 depending on how the scene is defined, and they can be converted into each other. When associating a keyword with a time interval of moving image data, it is convenient to select an expression method that is convenient for either displaying on the GUI or outputting as an XML file as metadata. It is also effective to convert the input data into a different representation method.

  Next, the processing table 113 according to the first embodiment will be described with reference to FIG. FIG. 6 is a configuration diagram of the processing table 113 according to the first embodiment.

  The processing table 113 includes a keyword 601, a processing content 602 for the keyword 601, a remark 603 for the keyword 601, a keyword ID 604 that is an identification number for identifying the keyword 601, and an identification number for identifying the processing content 602. A plurality of sets of process IDs 605 are stored.

  The processing table 113 may be described as array data used in a program such as C language, or may be data in the XML format as shown in FIG. 2. No need for ingenuity.

  Further, the processing content 602 associated with each keyword 601 may be one, plural, or zero.

  For example, the processing content “importance” 602a corresponds to the keyword “highlight” 601a. This means that the keyword 601a of highlight has an attribute of importance 602a, and as shown in FIG. 2, the highlight 601a is a word associated with a value 0.7 of importance Importance in the metadata. Indicates.

  For example, the keyword “day” or “night” is an absolute concept of whether or not, and has no level. For this reason, scenes associated with the keyword daytime have no level and are equally daytime. In other words, a statement that importance is given at noon and this scene is more applicable to noon is nonsense.

  On the other hand, the keyword 601a “highlight” is a relative concept having a level. When only highlight scenes of a moving image file are collected, each highlight scene has a level so that a highlight of an arbitrary length can be created according to the level of the scene. Therefore, the processing content called importance is associated with keywords having such relative levels.

  In the processing table 113 of FIG. 6, the keyword “goal” 601b is associated with a processing 602b called a front and rear context (longer front, shorter rear). This keyword 601b indicates that there are contexts before and after, and has a longer context on the front side and a shorter context on the rear side.

  More specifically, the goal scene of soccer is the soccer technology, even if the player's facial expression is interesting after the goal. Is less interested. In other words, if the length of the goal scene is extended, it is appropriate to extend the front side. However, as the time moves away from the goal moment, it becomes irrelevant from the goal scene, so it is better to extend the back side closer to the goal moment than the front side away from the goal moment. In this sense, the goal word has a longer front and a shorter back context.

  The keyword “free kick” 601c in FIG. 6 has a process 602c called context behind the goal 601b. This is because the free kick has more meaning after the moment of kick. Therefore, “free context” is associated with “free kick” 601c as processing 602c.

  Further, the process 602c is annotated with “curve c”, which indicates the shape of the process curve defined separately.

  As described above, the processing table 113 may describe not only a description of the processing method but also a link to a separately defined processing method. In this case, the convenience that the processing table does not need to be changed even when the processing method is changed is obtained.

  Note that the processing content 602 of the processing table 113 may be other than that shown in FIG. For example, the processing content 602 may be access control. As a result, it is possible to prevent a person who is not appropriate to acquire information on the keyword 601 from accessing this information.

  Here, an example of a processing method defined separately from the processing table 113 will be described with reference to FIGS. 7A to 7C. Fig.7 (a)-FIG.7 (c) are figures which show the example of the processing method defined separately. FIGS. 7A to 7C show the shapes of curves when the length of the scene is expanded or contracted.

  FIG. 7A shows a case where there are equal contexts before and after, and a new scene start and end time can be obtained by moving the threshold line up and down. The black band is the length of the scene described in the metadata, and the hatched band is the length of the expanded scene.

  FIG. 7B and FIG. 7C are examples of curves of the front context and the rear context, respectively, and scene expansion with a specific gravity before or after is performed by raising or lowering the threshold value.

  The definition of the process may be a curve as shown in FIG. When the curve c is designated in the processing table 113 in FIG. 6, the curve in FIG. 7C is referred to.

  In the processing table of FIG. 6, the keywords “Japan representative attack” 601d and “Osaka attack” 601e are associated with processing 602d and 602e “exclusive (team)”. The Japanese national team 601d and the Osaka 601e are soccer teams, respectively, and when they play, it means that when one is attacking, the other is defensive, and both are exclusive in the sense that they cannot attack at the same time.

  In the remarks column 603 of FIG. 6, other information related to the keyword 601 is described. In the example of FIG. 6, a URL is described in the remarks column 603, but it can be used as a place or name space in which more detailed instructions for the keyword 601 or the processing content 602 are described.

  Further, since the keyword 601 and the keyword ID 604 are associated with each other in the processing table 113, a description using the keyword ID 604 can be performed instead of the keyword 601.

  Further, since the processing content 602 and the processing ID 605 are associated with each other in the processing table 113, a description using the processing ID 605 can be performed instead of the processing content 602.

  As described above, the processing table 113 stores the keyword 601 and the processing content 602 adapted to the keyword 601 in association with each other. As a result, by using the processing table 113, processing that makes use of the characteristics of the words of the keyword 601 becomes possible.

  Next, a search request for video data search input from the request input unit 102 will be described with reference to FIG. FIG. 8 is a diagram illustrating a search request according to the first embodiment.

  In the first embodiment, the content data search system 100 inputs a search request from the request input unit 102 and transmits video data corresponding to the input search request.

  The search request 800 shown in FIG. 8 describes an item 801 to be searched, an input value 802, and a remark 803.

  Items to be searched 801 include a keyword 801a, a time 801b, a priority 801c, a charging method 801d, an image quality 801e, a transmission order 801f, and a source priority 801g.

  The keyword 801a is a part that specifies the content of the scene that the user wants to see.

  The time 801b is a part for designating the time that the user wants to send the video.

  The priority 801c indicates whether the scene length is shortened by giving priority to the time when the total time of the scene specified by the keyword is different from the specified time, or the scene length is given by giving priority to the scene. This is where you choose.

  For example, if “scene> time” is input to the item of priority 801c, a search result that does not change the length of the scene without considering the time is output (scene priority).

  The billing method 801d is a part for designating a method for paying the price for viewing the content. For example, the charging method 801d may have various forms such as credit card payment, payment as a call charge of a mobile phone company, and prepaid. In the example of FIG. 8, coupon payment is selected. Coupon payment provides video distribution such as points given to purchasers (so-called point cards) with product purchases, coupons according to the viewing time of advertisement video, coupons included in advertisements acquired by users, etc. The settlement is made by a kind of coupon presented to the user directly or indirectly by the person.

  The image quality 801e is a part for specifying the image quality such as the image compression rate, the number of frames per minute, the number of color gradations, and the number of pixels. Since the higher the image quality, the more necessary for streaming and download bandwidth, the user selects the image quality 801e according to the communication means and the state of the line. Alternatively, when the user is charged according to the amount of information to be streamed or downloaded, the user can specify the image quality with reference to the fee.

  The sending order 801f is a part for designating selection of whether the scenes are delivered in order of oldest time or in the order of importance associated with the scene when a plurality of scenes are delivered continuously. In the example of FIG. 8, since the transmission order 801f describes time series> importance, distribution is selected in the order of importance related to the scene.

  The source priority 801g is a part for designating what source is selected when content that matches the content requested by the user is provided from a plurality of sources. For example, the source priority 801g is input to give priority to a cheaper or free source when the source provision fee is different. In the example of FIG. 8, since the source priority 801g is described as free source priority, it is specified to give priority to the free source.

  Thus, the search request 800 input or transmitted by the user to the content data search system 100 of the first embodiment has a lot of information.

  Note that the items 801 and values 802 included in the search request 800 may be other than those shown in FIG. In addition, the search request 800 may not include all the items 801 illustrated in FIG.

  Next, the scene search unit 106 of the summary generation unit 105 inputs a search request 800, and video data search processing performed based on the input search request 800 will be described in detail with reference to FIG. FIG. 9 is a flowchart of the scene search algorithm of the scene search unit 106 according to the first embodiment.

  The processing from request input to scene transmission will be described in order with reference to FIG.

  First, the request input unit 102 of the communication processing unit 101 inputs a search request 800. Next, the request input unit 102 sends a search request 800 to the summary generation unit 105. On the other hand, the summary generation unit 105 receives and inputs the search request 800 in the scene search unit 106 (step 101).

  Next, the scene search unit 106 proceeds to a process of searching for a scene corresponding to the keyword specified in the item 801 of the search request 800 from the metadata.

  In the metadata, there is a description of a scene (segment) defining a certain time section (start time and end time) in the moving image file, and a keyword is associated with each scene.

  The scene search unit 106 extracts all scenes associated with the keyword specified in the search request 800 (step 102).

  However, if all scenes related to any of the keywords specified in the search request 800 are extracted, the time specified in the search request 800 may be exceeded.

  Therefore, the scene search unit 106 determines whether the total time of the scene extracted in step 102 exceeds the designated time 801b described in the search request 800 (step 103).

  If the total time of the scene extracted in step 102 does not exceed the specified time 801b, the scene search unit 106 does not perform any further search.

  On the other hand, if the total time of the scene extracted in step 102 exceeds the specified time 801b, the scene search unit 106 proceeds to a process for deleting the searched scene.

  First, when a plurality of keywords are specified in the search request 800, the scene search unit 106 searches for a scene associated with fewer specified keywords as a low priority scene.

  First, the scene search unit 106 extracts a scene having the lowest priority, that is, a scene associated with a small number of designated keywords. Next, the scene search unit 106 deletes the extracted lowest priority scene from the search candidate scene (step 104).

  For example, a scene (segment) associated with two keywords, the designated keywords “Osaka attack” and “highlight”, has a higher priority than a scene (segment) associated with only one of the keywords.

  In the case of metadata in which scene definitions do not overlap as shown in FIG. 3 and each scene can have a plurality of keyword associations, a high priority scene can be selected as described above, but as shown in FIG. If one scene has only one keyword relationship, a scene with high priority cannot be selected as described above. Therefore, the scene search unit 106 performs a search so as to extract the scene overlap time.

  For example, the time in the file of the scene associated with “Osaka attack” is 0: 1: 15-0: 2: 30, and the time in the file of the scene associated with “highlight” is 0: 2: If it was 00-0: 3: 30, it might be seen that the keywords “Osaka Attack” and “Highlight” were assigned at the overlap time of 0: 2: 0 to 0: 2: 30 it can. Therefore, the scene search unit 106 sets the scene of time 0: 2: 00 to 0: 2: 30 as a high priority scene.

  Next, the scene search unit 106 determines whether the total time of the scene after deleting the low priority scene in step 104 exceeds the specified time 801b described in the search request 800 (step 105).

  If it is determined in step 105 that the total scene time does not exceed the specified time 801b, the scene search unit 106 does not perform any further search.

  On the other hand, if it is determined in step 105 that the total scene time exceeds the specified time 801b, the scene search unit 106 returns to step 104 to search for and delete the next lowest priority scene. Then, the scene search unit 106 searches for a scene by repeating steps 104 and 105 so that the total time of the scene is close to the specified time.

  In this way, a scene matching the keyword is searched. When the priority is specified for each keyword in the search request, the ratio of the scene to be searched is adjusted according to a predetermined algorithm.

  In the first embodiment, the scene search unit 106 determines the priority of the scene by the number of designated keywords associated with the scene, but the priority of the scene may be determined by other methods.

  For example, the scene search unit 106 may determine the priority of the scene based on the importance of the scene described in the metadata. For example, 0.7 is described as the importance value of the scene in the portion indicated by 204 in FIG. The scene search unit 106 can prioritize scenes by comparing the importance values. In this way, the scene search unit 106 may determine the priority of the scene.

  In addition, although two examples of how the scene search unit 106 determines the priority of the scene have been given, these two methods may be used in combination using a specific calculation formula. As a result, there may be a case where a result with a more appropriate priority of the scene is obtained.

  By the way, by the above-described processing, a scene that matches the search request can be extracted so as not to exceed the specified time 801b. However, depending on the length of the scene described in the metadata, the difference between the extracted scene total time and the specified time may be different. Sometimes it ca n’t be ignored.

  Therefore, in order to solve this problem, in the first embodiment, the data management unit 107 next performs an algorithm for time optimization processing. This process is a process for expanding and contracting the length of the scene extracted using the process table 113 shown in FIG.

  First, the data management unit 107 lists the keywords associated with the scenes extracted by the scene search unit 106 in the process up to step 105 (step 106).

  Next, the data management unit 107 searches for the same keyword as the enumerated keyword in the context processing designation keywords described in the processing table 113 (step 107).

  Since the total time of the scenes extracted in the processes up to step 105 is shorter than the specified time, the data management unit 107 next shifts to a process of decompressing the scene so that the total time of the extracted scenes matches the specified time.

  First, it is generally desirable for the data management unit 107 to expand a scene having a lot of important contents, that is, a scene with higher importance. Therefore, the data management unit 107 selects a scene assigned with a higher importance from the scenes assigned with the keyword retrieved in step 107 (step 108).

  Next, the data management unit 107 refers to the processing table 113 and checks the context instruction. Then, the data management unit 107 expands the scene selected in step 108 in accordance with the context instruction described in the processing table 113 (step 109).

  For example, if the keyword is a free kick in FIG. 6, the data management unit 107 refers to FIG. 7C because the processing content is the post-context (curve C). Next, the data management unit 107 assigns the start and end times of the scene before and after the black rectangle in FIG. 7C, and sets the threshold line and the length of the hatched rectangle as the target time. Move up and down.

  Next, the data management unit 107 acquires a start time and an end time corresponding to the hatched rectangle. Then, the data management unit 107 acquires the acquired video data 109 and 111 corresponding to the start time and end time. That is, the data management unit 107 acquires the scenes before and after the currently processed scene, and performs processing for connecting to the currently processed scene.

  The target time is a difference between the specified time and the total time. However, when adjusting the time of a plurality of scenes (segments), the adjustment time may be allocated at a predetermined ratio.

  Next, the data management unit 107 determines whether or not the total time of the scene after the scene expansion processing in step 109 exceeds the specified time 801b of the search request 800 (step 110).

  Then, the data management unit 107 repeats the processing of step 109 and step 110 until the total scene time exceeds the specified time 801b of the search request 800.

  In the loops of steps 109 and 110, if the length a that is extended at a time is increased, the loop can be exited with a small number of loops, and the processing time can be shortened. There is a possibility of becoming large.

  Therefore, in the first embodiment, fine adjustment of the scene length is performed.

  Specifically, first, the data management unit 107 determines whether or not the excess time with respect to the specified time of the total time of the scene processed in the processing up to step 110 is within an allowable range (step 111).

  If it is determined in step 111 that the excess time of the scene exceeds the allowable range, the data management unit 107 shortens the scene length by using the length b that is smaller than the length a that is used in step 109. (Step 112). Then, the data management unit 107 repeats the processing of step 111 and step 112 until the scene excess time falls within the allowable range.

  The scene searched in this way is shown in FIG. FIG. 10 is a diagram for explaining processing when the data management unit 107 searches for a highlight scene in a specified time of 1 minute.

  In FIG. 10, since the importance level of the fourth scene 301d is 1.0 but only 40 seconds, the data management unit 107 also selects the third scene 301c whose importance level is 0.8.

  However, since the total time of the scene 301c and the scene 301d exceeds 60 seconds, the data management unit 107 expands and contracts the length of the third scene 301c according to the instruction of the processing table 113, and sets a section where the total time is 60 seconds. It is a search result.

  In this way, the data management unit 107 searches for a scene corresponding to the search request 800 using the processing table 113, and adjusts the total time of the searched scene to match the specified time. In this way, the data management unit 107 can expand the scene so that the features of the scene can be made clearer by expanding the scene by the expansion method according to the characteristics of the keyword assigned to the scene.

  Then, the data management unit 107 sends the searched scenes to the video data sending unit 103 in the order of the transmission order 801f described in the search request 800, and the video data sending unit 103 sends the scenes sent sequentially ( Step 113).

  As described above, according to the first embodiment, content data is managed by the metadata and the processing table 113, so that searches and summaries can be made using the characteristics of the keywords used in the metadata. Can be realized.

  Further, according to the first embodiment, if the content of the processing table 113 is changed, the content of the processing can be arbitrarily changed without changing the content data search system 100. Thereby, convenience is high when processing is changed without changing the content data search system 100 itself. If a plurality of processing tables 113 are prepared and the optimum processing table 113 is selected according to the situation, the optimum search can be realized according to the situation.

  In the first embodiment, an example of only context processing in the processing table 113 has been described, but other processing instructions may be referred to the processing table 113 in the same manner.

  For example, in the case of the processing content “exclusive”, it is possible to check whether there is a contradiction in the keyword specification of the search request.

  More specifically, it is assumed that an Osaka attack and a Tokyo attack are designated as the keyword 801a in FIG. Osaka and Tokyo are soccer team names, respectively, and the processing contents are designated as “exclusive (team)” by 601e in FIG. This indicates that each keyword is exclusive and does not hold at the same time. Therefore, unless the rules are revised so that soccer uses two soccer balls, there is no scene that satisfies both the Osaka attack and the Tokyo attack, and the search request cannot be answered. In such a case, if the validity of the search request is checked first with reference to the processing table, the request request creator can be pointed out to the search request creator without performing a useless search, thereby improving the search efficiency. .

  For example, in the case of the processing content of “importance”, even if the importance value is not assigned to the keyword, the processing can be performed assuming that a value such as 0.5 as a reference value is assigned. In this case, the convenience of scene evaluation is increased.

  In addition, when the search target is spread over multiple files, control the numerical value by biasing the importance of the scene (segment) or putting on clogs according to the priority of the file. It is also possible and highly convenient.

  In addition, when the method of assigning keywords to each of a plurality of files and how to determine the importance value are different, the importance level of a similar scene is 0.7 in one file and 1.0 in another file. In this case, the value given to the entire file can be corrected in order to ensure consistency of values.

  In the algorithm shown in FIG. 9, the process is repeated until the total time of the scene falls within the specified time. However, another method may be used to match the specified time. For example, if a difference or ratio with respect to the specified time is calculated, and processing is performed so that the corrected time coincides with the specified time at once, the processing loop becomes unnecessary.

  In the algorithm shown in FIG. 9, the case where the total time exceeds the specified time has been described. However, if the total time is shorter than the specified time, the reverse process may be performed so that the total time exceeds the specified time.

  When the scene time is expanded or contracted so as to coincide with the designated time, it may not be appropriate to expand or contract the time so as to be exactly the same for reasons such as the processing time and the balance of the scene configuration. In such a case, a slight error may be left as long as it is possible to enter during a certain period of time.

  In addition to online distribution using communication means such as a network, content data, video data, and video data can be distributed from creators to viewers through offline distribution that is held and distributed on portable media such as DVD disks and flash memory. The form passed to In this case, when content data is recorded on the medium, the metadata and the processing table are recorded on the medium and simultaneously distributed. This has the following effects.

  First, in comparison with only content data and metadata, detailed instructions for each keyword described in the processing table 113 as described above can be reflected at the time of retrieval. This makes the search result closer to the result desired by the user.

  Second, the processing table 113 can use one file in common for each metadata. Thereby, it is not necessary to have the processing table 113 for each metadata, and the capacity of the medium can be saved.

  Note that the processing performed by the content data search system 100 may be a program and a general-purpose computer may execute the program.

  In the first embodiment, the processing content for the keyword is described as being owned by the processing table 113. However, the processing content for the keyword may be directly written in the metadata.

(Embodiment 2)
In the second embodiment of the present invention, an authoring system that creates metadata describing the content data will be described.

  First, the configuration of the authoring system according to the second embodiment will be described with reference to FIG. FIG. 11 is a configuration diagram of an authoring system according to the second embodiment.

  An authoring system 1300 according to the second embodiment employs a configuration in which an input / output processing unit (GUI) 1301, an authoring processing unit 1305, and a content data providing unit 1308 are connected by a data bus 1304.

  The input / output processing unit 1301 is provided with an operation input unit 1302 and a display unit 1303.

  The operation input unit 1302 inputs a user operation and sends the input user operation to the display unit 1303, the authoring processing unit 1305, and the content data providing unit 1308. The operation input unit 1302 is an input device such as a keyboard, a mouse, or a touch panel, and this input / output processing is a so-called GUI. The operation input includes not only a user input but also an automatic processing input such as voice recognition.

  The display unit 1303 displays the status operated by the user.

  The data bus 1304 is a bus for transferring data.

  The content data providing unit 1308 includes a video data storage unit 114 that stores and manages a plurality of video data 109 and 111. The content data providing unit 1308 also includes a metadata storage unit 1313 that stores and manages the metadata 1310 and 1312 created by the authoring processing unit 1305. In the metadata 1310 and 1312 created by the authoring processing unit 1305, information as described in the first embodiment is described.

  In the metadata 1310 and 1312, a link as to which video data 109 and 111 is the metadata is described, and the video data 109 and 111 and the metadata 1310 and 1312 are associated with each other.

  Note that the video data 109 and 111 and the metadata 1310 and 1312 are often one-to-one, but this is not necessarily the case, and one metadata may hold information related to a plurality of video data, There may be a plurality of metadata related to one video data.

  In addition, the content data providing unit 1308 includes a table storage unit 116 that stores and manages the processing table 133. The processing table 133 is a correspondence table between keywords used in the metadata 1310 and 1312 and processing.

  As the content data providing unit 1308, an information holding medium such as a hard disk, an optical disk, or a magnetic disk can be used.

  The video data 109 and 111 are digital files using an encoding format such as MPEG.

  The authoring processing unit 1305 is provided with a moving image access unit 1306 and a data creation unit 1307.

  The moving image access unit 1306 reads out the video data 109 and 111 in accordance with an operation input from the user and displays it on the display unit 1303. The user performs an operation of assigning a keyword to a desired section of the video data 109 and 111 in the operation input unit 1302, which is received by the data creation unit 1307, creates metadata 1310 and 1312, and creates a content data provision unit 1308. Keep in.

  The authoring system 1300 according to the second embodiment includes a plurality of means as shown in FIG. 11, but these can be realized by a single personal computer or a combination of a plurality of devices.

  For example, it is possible to store the content data providing unit 1308 in a network-connected storage, perform processing by the server, and perform input / output to / from the user with a personal computer.

  In addition, the authoring system software can be used in various forms such as a device built-in, a WEB service via a network, and retention on a disk medium.

  Next, the GUI of the authoring system 1300 according to the second embodiment will be described with reference to FIG. FIG. 12 is a diagram illustrating an example of the GUI of the authoring system 1300 according to the second embodiment.

  A GUI operation screen 1400 is displayed on the display unit 1303. In order to input the user's intention using the operation screen 1400, an operation input unit 1302 which is an input means such as a mouse or a keyboard is used.

  A menu unit 1401 is provided on the upper left side of the operation screen 1400. The menu 1401 portion includes a menu 1401a and a menu button 1401b, and is used when selecting and saving a file.

  A moving image display unit 1402 is provided slightly above the center of the operation screen 1400. The moving image display unit 1402 displays the image of the moving image file selected by the user.

  Further, a file information button 1405 is provided on the right side of the moving image display unit 1402 on the operation screen 1400. A file information selection button 1405 is used when a moving image file to be displayed on the moving image display unit 1402 is selected.

  A time axis 1408 is provided substantially at the center of the operation screen 1400. The time axis 1408 indicates which time of the moving image file currently displayed on the moving image display unit 1402 is reproduced by the marker 1408a.

  Further, the user can seek the video by moving the marker 1408a.

  A keyword selection button 1406 is provided on the right side of the moving image display unit 1402 on the operation screen 1400. The keyword selection button 1406 is used for displaying a keyword list and selecting a keyword to be used for authoring from the displayed keyword list.

  Further, a plurality of keyword buttons 1407 are provided near the lower portion of the keyword selection button 1406 on the operation screen 1400. The keyword button 1407 is a button corresponding to the keyword selected by the keyword selection button 1406.

  A list 1411 is provided at the lower left end of the operation screen 1400. A list 1411 is a list in which keywords corresponding to the keyword button 1407 are listed.

  A plurality of scenes (segments) 1412 of the moving image file displayed on the moving image display unit 1402 are arranged in the order of time series in the vicinity below the time axis 1408 of the operation screen 1400. As shown in FIG. 3, the scenes are continuous and do not overlap.

  The scene 1412 is a button, and has a function of seeking a video at the start time of the scene when pressed.

  In addition, a scene division button 1403 is provided in the vicinity of the right side of the moving image display unit 1402 on the operation screen 1400. A scene division button 1403 is used when a scene is divided.

  A plurality of video seek buttons 1409 are provided at the upper left end of the operation screen 1400. A video seek button 1409 performs video playback and seek before and after. Further, when the scene division button 1403 is pressed while the moving image file is being played with the video seek button 1409, the position (time) becomes a break of the scene.

  When a moving image file is read when authoring for the first time, the start time to the end time of the entire moving image file are defined as one scene, and are redefined as a plurality of scenes by a scene division button 1403.

  Further, a scene combination button 1404 is provided near the right side of the scene division button 1403 on the operation screen 1400. The scene combination button 1404 is used to combine a scene including the current position of the moving image file with either the previous or next scene.

  The user can freely define a scene (segment) by operating the scene division button 1403 and the scene combination button 1404.

  When a video file is played or seeked using the video seek button 1409 and the keyword button 1407 is pressed at that position, a keyword is assigned to the scene including the time, and the time is below the time axis 1408 of the operation screen 1401 and A tag button 1413 is displayed at a corresponding position on the axis 1408. The tag button 1413 is arranged in the same column as the keyword list 1411 so that it can be understood which keyword is associated. The shade of the color of the tag button 1413 represents the strength of the importance, and the importance can be changed by clicking the tag button 1413 or the like.

  In addition, a scene display button 1410 is provided near the top of the keyword list 1411 of the operation screen 1400. A scene display button 1410 is pressed to display a list of thumbnail images of each scene.

  As described above, this GUI makes it possible to divide a moving image file into arbitrary scenes and to assign arbitrary keywords to the divided scenes.

  In addition, scene definitions and keyword associations may be written in the content data providing unit 1308 as metadata for each operation, but are usually held in the memory of the authoring processing unit 1305 and collectively the content data at the end of the authoring operation. Write to the providing unit 1308.

  Next, an algorithm of the authoring system 1300 according to the second embodiment will be described with reference to FIG. FIG. 13 is an operation flowchart of the authoring system 1300 according to the second embodiment.

  When there is an input to select a moving image file from the operation input unit 1302 of the input / output processing unit 1301, the moving image access unit 1306 of the authoring processing unit 1305 accesses the content data providing unit 1308 and is selected from the operation input unit 1302. Movie files (video files) 109 and 111 are read (step 1301).

  Next, the moving image access unit 1306 displays the read moving image file on the display unit 1303 of the input / output unit (GUI) 1301, thereby presenting it to the user (step 1302).

  On the other hand, the user operates the video seek button 1409 to play a video or seek a video, and further operates the keyword selection button 1406 to associate a keyword with a desired scene or operate the tag button 1413. To set the importance.

  In response to this, the operation input unit 1302 sends information related to moving image playback and moving image seeking, information such as a keyword selected by the user and importance, to the data creation unit 1307 (step 1303).

  Then, the data creation unit 1307 associates the keyword selected by the user with the scene selected by the user and sets the importance (step 1304).

  Next, the data creation unit 1307 searches whether the keyword selected by the user is defined in the processing table 133 (step 1305).

  If there is a keyword that matches the processing table 133, the data creation unit 1307 performs the processing content specified in the processing table 133 (step 1306).

  Specifically, for example, if the processing content “exclusive (team)” is defined in the processing table 133, the data creation unit 1307 automatically selects the keyword of the battle team where the keyword is not associated. Associate. Further, the data creation unit 1307 acquires the battle team from information such as file information given to the entire file. This eliminates the need for the user to assign an exclusive keyword, that is, a keyword related to the battle team, and has the advantage of reducing the amount of authoring work.

  In addition, since it is contradictory that exclusive keywords are associated with the same time interval (scene), it is convenient to prohibit such contradictory operations in the authoring operation.

  Further, for example, if the processing content “importance” is defined in the processing table 133, the data creation unit 1307 automatically associates the default value without the input of the importance from the user. By automatically assigning importance in this way, the user's workload is reduced.

  Further, the default value of importance is a value set in advance by the user. For example, when the importance range is from 0 to 1, 0.5, which is the midpoint, may be set, or another value may be used.

  If the contents of the file are important in all time zones, it is also effective to add an offset value and set a higher value as the default importance value than other files. This is because when content that is boring as a whole and very interesting content are searched simultaneously, interesting content should be searched preferentially, and this can be achieved by increasing the importance offset value. is there.

  For example, if the process “133” is defined in the processing table 133, the data creation unit 1307 associates the importance level with the preceding and following scenes step by step when the keyword and the importance level are input from the user. I do. In this way, by automatically assigning importance to the previous and next scenes that are currently being processed, it is not necessary for the user to assign importance to the previous and subsequent scenes that are currently being processed. Very convenient.

  An example of how to assign this importance will be described with reference to FIGS. 14 (a) and 14 (b).

  FIG. 14A is a diagram illustrating keywords and importance levels input to a scene by a user operation.

  In the example of FIG. 14A, the keyword “goal” and the importance “0.8” are associated with a section having a start time of 0:30:30 and an end time of 0:32:30.

  The scene, keyword, and importance state shown in FIG. 14A corresponds to the state in step 1304 shown in FIG.

  In the next step 1305, the data creation unit 1307 searches the processing table 133.

  In the processing table 133, “front and back context (longer front, shorter rear)” is defined as the processing content for the keyword “goal”.

  Therefore, the data creation unit 1307 adds importance to the two scenes in the front and the one scene in the back in stages. FIG. 14B shows a scene to which importance is automatically added by the data creation unit 1307.

  The keyword “goal” is focused on before. For this reason, as can be seen from FIG. 14B, the data creation unit 1307 increases the importance of the scene before the subsequent scene.

  In the example of FIG. 14B, importance is automatically added to a total of three scenes. However, the number of scenes to which importance is added, the value of importance, and the length of each time are appropriate for the authoring system. You can decide.

  Further, detailed setting instructions such as the number of scenes to which importance is added, the value of importance, and the length of time may be described in the processing table 133 or the link destination in the remarks column described in the processing table 133. According to this, it is possible to select a process according to the keyword more finely.

  In this way, the data creation unit 1307 updates the scene information of the metadata (step 1307).

  Then, the data creation unit 1307 repeats the loop of Steps 1303 to 1307 until the keyword association authoring operation by the user is completed (Step 1308).

  Then, when the authoring operation for keyword association by the user ends and the user inputs to save the generated metadata (step 1309), the data creation unit 1307 stores the meta data as a file in the content data providing unit 1308. Data 1310 and 1312 are stored (step 1310).

  On the other hand, if the user does not input to save the generated metadata, the data creation unit 1307 does not create metadata in the content data providing unit 1308.

  For example, if the user reads and edits the existing metadata, the metadata of the content data providing unit 1308 is not updated unless saving is selected in step 1309.

  As described above, according to the authoring system of the second embodiment, authoring processing using the processing table 133 can be performed. Thereby, it is possible to support the user's keyword assignment operation, and work efficiency can be significantly improved.

  Further, according to the second embodiment, if the content of the processing table 133 is changed, the processing content can be changed to an arbitrary one without changing the authoring system at all. Therefore, the processing can be performed without changing the authoring system itself. Convenient to change.

  If a plurality of processing tables are prepared and the optimum processing table is selected according to the situation, the optimum automatic authoring can be realized according to the situation.

(Embodiment 3)
The third embodiment is an improvement of the second embodiment, in which the amount of information conveyed by the metadata is maintained and the metadata is not increased.

  Specifically, in the second embodiment, information is added to the metadata according to the processing content defined in the processing table. However, in the third embodiment, the ID of the processing defined in the processing table is used as a keyword of the metadata. Associate with.

  Hereinafter, the algorithm of the authoring system according to the third embodiment will be described with reference to FIG. FIG. 15 is an operation flowchart of the authoring system 1300 according to the third embodiment. Note that the configuration of the authoring system according to the third embodiment is the same as that of the second embodiment, and a description thereof will be omitted.

  When there is an input to select a moving image file from the operation input unit 1302 of the input / output processing unit 1301, the moving image access unit 1306 of the authoring processing unit 1305 accesses the content data providing unit 1308 and is selected from the operation input unit 1302. Movie files (video files) 109 and 111 are read (step 1501).

  Next, the moving image access unit 1306 displays the read moving image file on the display unit 1303 of the input / output unit (GUI) 1301, thereby presenting it to the user (step 1502).

  On the other hand, the user operates the video seek button 1409 to play a video or seek a video, and further operates the keyword selection button 1406 to associate a keyword with a desired scene or operate the tag button 1413. To set the importance.

  In response to this, the operation input unit 1302 sends information related to the reproduction of the moving image and the seeking of the moving image, information such as the keyword selected by the user and importance, to the data creation unit 1307 (step 1503).

  Then, the data creation unit 1307 associates the keyword selected by the user with the scene selected by the user and sets the importance (step 1504).

  Next, the data creation unit 1307 searches whether the keyword selected by the user is defined in the processing table 133 (step 1505).

  If there is a matching keyword in the processing table 133, the data creation unit 1307 performs processing for describing the processing ID specified in the processing table 133 in the metadata (step 1506).

  Specifically, for example, for the keyword “goal”, the data creation unit 1307 describes the process ID 3 corresponding to “goal” in the process table 113 of FIG. 6 in the metadata. The process ID 605 is a number assigned to specify the process, and the process content 602 associated with the keyword can be specified by using the correspondence table of the process ID 605 and the process content 602.

  An example in which the process ID 602 is described in the metadata is shown in FIG. In the part indicated by 1801 in FIG. 16, a process ID 602 is described as one of the attribute values of XML.

  In addition to this method, the method of describing the process ID 602 in the metadata is a method of describing it as XML NODE or a method of describing “goal: process ID: 3” as a part of the keyword value. Etc.

  The same effect can be obtained by describing the keyword ID 604 of FIG. 6 in the metadata instead of the process ID 602.

  In this way, the data creation unit 1307 updates the scene information of the metadata (step 1507).

  Then, the data creation unit 1307 repeats the loop from step 1503 to step 1507 until the keyword association authoring operation by the user is completed (step 1508).

  Then, when the authoring operation for keyword association by the user ends and the user inputs to save the generated metadata (step 1509), the data creation unit 1307 stores the meta data as a file in the content data providing unit 1308. Data 1310 and 1312 are stored (step 1510).

  As described above, according to the third embodiment, it is possible to describe the ID corresponding to the processing content instead of describing the processing content itself in the metadata. The metadata described by such a method can be smaller in file size than the metadata created in the second embodiment.

  In addition, the search system that uses the metadata generated in the third embodiment uses the metadata and the processing table (or the processing ID and the processing correspondence table) together, and uses the metadata in the second embodiment. What is necessary is just to perform the automatic assignment | providing of data as demonstrated and to restore | restore the metadata of Embodiment 3 to the metadata produced in Embodiment 2. FIG. Thereby, the same effect as described in the second embodiment can be obtained.

  The difference between the second embodiment and the third embodiment is whether the extended interpretation specified in the metadata processing table is performed at the time of authoring or when the metadata is used. There is no difference in effect, and the processing load is reduced. It should be selected from the viewpoint of which one is more advantageous and whether the size of the metadata needs to be reduced.

  As described above, according to the present invention, the processing method of content data can be set in detail by using the processing table for the content data. As a result, when a scene associated with a specific keyword is searched and extracted from the content data, the total time of the search result can be finely adjusted to a desired time.

1 is a configuration diagram of a content data search system according to a first embodiment of the present invention. The figure which shows the specific example of the metadata concerning Embodiment 1. The figure which shows typically the provision state of the keyword in the metadata concerning Embodiment 1. FIG. The figure which shows typically the another provision state of the keyword in the metadata concerning Embodiment 1. FIG. The figure which shows typically the another provision state of the keyword in the metadata concerning Embodiment 1. FIG. Configuration diagram of processing table according to the first embodiment (A) 1st figure which shows the example of the processing method concerning Embodiment 1, (b) 2nd figure which shows the example of the processing method concerning Embodiment 1, (c) The process concerning Embodiment 1 Third diagram showing an example of the method, The figure which shows the search request concerning Embodiment 1. Flowchart of a scene search algorithm of the scene search unit according to the first embodiment The figure for demonstrating the process in case the data processing part concerning Embodiment 1 searches a highlight scene in designated time 1 minute. Configuration diagram of authoring system according to Embodiment 2 of the present invention FIG. 6 is a diagram illustrating an example of a GUI of an authoring system according to the second embodiment. Operation Flow of Authoring System According to Second Embodiment (A) The figure which shows the keyword and importance which were input with respect to the scene by user operation, (b) The figure which shows the scene which the data preparation part concerning Embodiment 2 added importance automatically Flow chart of operation of authoring system according to third embodiment of the present invention The figure which shows the metadata concerning Embodiment 3.

Explanation of symbols

DESCRIPTION OF SYMBOLS 100 Content data search system 101 Communication processing part 102 Request input part 103 Video | video data transmission part 105 Summary production | generation part 106 Scene search part 107 Data management part 108, 1308 Content data provision part 114 Video | video data storage part 115, 1313 Metadata storage part 116 Table storage unit 1300 Authoring system 1301 Input / output processing unit 1302 Operation input unit 1303 Display unit 1305 Authoring processing unit 1306 Video access unit 1307 Data creation unit

Claims (18)

  A metadata reference means for referring to metadata describing contents of content data, and a table reference means for referring to a processing table describing processing contents corresponding to a keyword used in the metadata are provided. Content management device.   The processing table provision apparatus which provides the said processing table with respect to the content management apparatus of Claim 1.   Content data reference means for referring to content data composed of a plurality of scenes, metadata reference means for referring to metadata in which the contents of the scene are described by keywords, and processing contents corresponding to the keywords used in the metadata are described A table reference means for referring to the processed table, a scene search means for inputting a search keyword and searching for the scene to which the search keyword is assigned from the content data reference means, and provided to the searched scene. Data processing means for extracting the processing content corresponding to the keyword from the processing table and performing a processing corresponding to the extracted processing content on the searched scene. Search device.   The processing content is a process of expanding the scene, and the data processing means performs the search so that the total time becomes the specified time when the total time of the searched scene is less than a desired specified time. 4. The content data search apparatus according to claim 3, wherein the scene is expanded.   The process of expanding the scene is a process of adding a part of the scene before or after the searched scene to the searched scene according to the characteristics of the keyword assigned to the scene. The content data search device according to claim 4.   6. The content data according to claim 4, wherein importance is given to the scene, and the data processing means expands the scene having the highest importance among the searched scenes. Search device.   Content data reference means for referring to content data composed of a plurality of scenes, table reference means for referring to a processing table describing processing details corresponding to a predetermined keyword, and a keyword assigned to the scene are input and input The processing contents corresponding to the keyword is extracted from the processing table, and a description corresponding to the extracted processing contents is written to the scene, thereby generating data indicating the metadata of the contents data. And an authoring processing apparatus.   The processing content is exclusive processing, and the data creation means assigns a keyword having a meaning exclusive to the input keyword to a scene other than the scene to which the input keyword is assigned. The authoring processing apparatus according to claim 7.   8. The authoring processing apparatus according to claim 7, wherein the processing content is an importance level assigning process, and the data creating unit assigns a predetermined importance level to the scene to which the input keyword is assigned.   The process content is a process of assigning importance levels in stages, and the data creation means sets the importance level of the scene to which the keyword is assigned to the scene before and after the scene to which the input keyword is assigned. 8. The authoring processing apparatus according to claim 7, wherein importance is given step by step.   The processing content is processing for giving importance with emphasis on the front, and the data creation means is behind the scene to which the keyword is assigned to the scene before the scene to which the input keyword is assigned. The authoring processing apparatus according to claim 10, wherein an importance level higher than that of a scene is given.   The processing content is a process of giving importance with emphasis on the back, and the data creation means is arranged before the scene to which the keyword is assigned to the scene behind the scene to which the input keyword is assigned. The authoring processing apparatus according to claim 10, wherein an importance level higher than that of a scene is given.   Content data reference means for referring to content data composed of a plurality of scenes, table reference means for referring to a processing table describing processing details corresponding to a predetermined keyword and identification information for the processing details, and giving to the scene The identification information corresponding to the input keyword is extracted from the processing table, and the extracted identification information is described in the scene to generate metadata indicating the content data. An authoring processing apparatus comprising: a data creating unit;   A content management method characterized in that the content data is managed using metadata describing the content of the content data and a processing table describing a processing method corresponding to a keyword used in the metadata.   A step of referring to content data comprising a plurality of scenes, a step of referring to metadata describing the contents of the scene with keywords, and a processing table describing processing details corresponding to the keywords used in the metadata A step of inputting a search keyword and searching for the scene to which the search keyword is assigned from the content data, and the processing content corresponding to the keyword assigned to the searched scene. A content data search method comprising: extracting from a table; and performing a process corresponding to the extracted processing content on the detected scene.   A step of referring to content data comprising a plurality of scenes; a step of referring to a processing table describing processing details corresponding to a predetermined keyword; a step of inputting a keyword to be given to the scene; and the input keyword Extracting the processing content corresponding to the processing table from the processing table, generating the metadata corresponding to the content data by performing a description corresponding to the extracted processing content on the scene, An authoring processing method characterized by comprising:   A step of referring to content data comprising a plurality of scenes on a computer, a step of referring to metadata describing the contents of the scene with keywords, and a processing table describing processing details corresponding to the keywords used in the metadata , A step of inputting a search keyword and searching the scene to which the search keyword is assigned from the content data, and the processing content corresponding to the keyword assigned to the searched scene And a step of performing, on the detected scene, a process corresponding to the extracted processing content.   A step of referring to content data comprising a plurality of scenes, a step of referring to a processing table describing processing details corresponding to a predetermined keyword, a step of inputting a keyword to be assigned to the scene, and an input; The process contents corresponding to the keyword are extracted from the process table, and a description corresponding to the extracted process contents is performed on the scene to generate metadata indicating the contents data contents. And a step.

Images