JP5213797B2 - Metadata search device, metadata search method, metadata search program, and metadata search system - Google Patents

Description

  The present invention relates to a metadata search device, a metadata search method, a metadata search program, and a metadata search system.

  Various accompanying information related to video content and music content is assigned as metadata to the content. Taking song content as an example, a song name, artist name, genre, and the like are given to the content as metadata. Furthermore, not only metadata for the entire content but also text information of lyrics accompanying the singing video every hour, performer information accompanying the drama video every hour, etc. Is added to a predetermined section or region of the content.

  As a technique for searching for this metadata, there is one described in Japanese Patent Application Laid-Open No. 2008-219608. In the display device disclosed in Patent Document 1, a reference time that is an extraction start time of feature data of a playback content is acquired, and feature data is extracted from the playback content. Based on the feature data, the feature data in the stored content is searched, the time in the stored content corresponding to the feature data derived from the playback content is specified as the corresponding reference time, and the current playback time of the playback content Is acquired. Then, a playback corresponding reference time is calculated based on the playback time, the reference time, and the corresponding reference time, and corresponds to the time corresponding to the playback corresponding reference time from the metadata associated with each time in the stored content. The attached metadata is selected and displayed.

JP 2008-219608 A

  However, in the display device described in Patent Document 1, as a result of search based on feature data of reproduced content, metadata of a plurality of accumulated contents may be extracted. In particular, when a search is performed in a longer playback content section, feature data of the stored content that is similar to part of the feature data in this section increases, and more metadata of the stored content is extracted. When a plurality of metadata is extracted, the display device described in Patent Document 1 does not have a standard for which metadata should be displayed. Therefore, the metadata cannot be displayed, or all metadata is displayed. It will be displayed uniformly. For this reason, the display device described in Patent Document 1 has a drawback that it cannot select metadata that matches the playback content, or that it takes time for a user or the like to select metadata.

  The present invention has been made in view of the above points, and an object of the present invention is to provide a metadata search device, a metadata search method, a metadata search program, and a metadata search program that can quickly select metadata that matches content. To provide a data retrieval system.

  The present invention has been made in order to solve the above-described problems. The present invention is a metadata search apparatus that searches for stored content of a signal similar to the content signal and outputs metadata of the searched stored content. Search for outputting metadata of the stored content section or region of a signal similar to a signal included in the content section or region based on storage section length information indicating the size of the stored content section or region A metadata search device comprising a result output unit.

  The present invention also provides a metadata search apparatus that searches for stored content of a signal similar to a content signal and outputs metadata of the searched stored content, and includes information on a query section indicating the section of the content or the area. , A section length correspondence storage unit that prestores section length correspondence information that associates the section of the stored content of the signal similar to the signal included in the query section or the storage section length information indicating the size of the area. A section length correspondence acquisition unit to be read from, a search command acquisition unit for acquiring information on a search section indicating the section or area of the content, and a section or area of the stored content of a signal similar to a signal included in the search section A metadata search unit for extracting the metadata of the section, and storage section length information of the section length correspondence information, the search Based on the accumulated segment length information corresponding to the query section included between a metadata search apparatus characterized by comprising, said metadata search unit outputs the extracted metadata retrieval result output unit.

  In the metadata search device according to the present invention, the search result output unit is storage section length information of the section length correspondence information, and storage section length information corresponding to a query section included in the search section. The metadata extracted by the metadata search unit is output in descending order of the size of the section or area indicated by.

  According to the present invention, in the metadata search apparatus, the storage section length information is information on a storage section indicating a section or a region of the stored content of a signal similar to a signal included in the query section. Features.

  In the metadata search device according to the present invention, the search result output unit is storage section length information of the section length correspondence information, and storage section length information corresponding to a query section included in the search section. Is output.

  In the metadata search device according to the present invention, the search result output unit calculates a histogram representing the appearance frequency of stored content for each content section or region in the search section. The metadata extracted by the metadata search unit is output based on the appearance frequency shown.

  Further, the present invention is characterized in that, in the metadata search device, the output order of metadata extracted by the metadata search unit is determined based on the appearance frequency indicated by the calculated histogram.

  In the metadata search device, the present invention provides a content correspondence acquisition unit that acquires a plurality of correspondences including the query interval and the accumulation interval of a signal similar to the content signal of the query interval. , A combined correspondence generation unit that generates a combined correspondence by combining a plurality of correspondences between the query interval and the storage interval, the query interval combined by the combined correspondence generation unit, and the combined correspondence relationship And a section length correspondence generation unit that stores storage section length information indicating the size of the storage section in the section length correspondence storage unit as the section length correspondence information.

  Further, the present invention provides a metadata search method in a metadata search apparatus that searches for stored content of a signal similar to a content signal and outputs metadata of the searched stored content. Corresponding to the section length, the information of the query section indicating the section or area of the content and the storage section length information indicating the size of the stored content section or area of the signal similar to the signal included in the query section are associated. A section length correspondence acquisition process for reading the relation information in advance from a section length correspondence storage unit, a search command acquisition unit for acquiring information on a search section indicating the section or area of the content, A data search unit stores a section or region of the stored content of a signal similar to the signal included in the search section. A metadata search process for extracting data, and a search result output unit, based on storage section length information corresponding to a query section included in the search section, which is storage section length information of the section length correspondence information, A metadata search method comprising: a search result output process for outputting metadata extracted in the metadata search process.

  In addition, the present invention searches for a stored content of a signal similar to a content signal, and outputs a metadata of the searched stored content to a computer of a metadata search apparatus, and information on a query interval indicating the content interval or the region. And a section length correspondence storage that prestores section length correspondence information that correlates the section of the stored content of the signal similar to the signal included in the query section or the storage section length information indicating the size of the area. Section length correspondence acquisition means for reading out from the section, search command acquisition means for acquiring information on the search section indicating the content section or area, and the stored content section or area of the signal similar to the signal included in the search section Metadata search means for extracting metadata, and storage section length information of the section length correspondence information. Based on the accumulated segment length information corresponding to the query sections included in the search interval, which is the metadata retrieval program metadata search means search result output means for outputting the extracted metadata in, to function as a.

  Further, the present invention is a metadata management system comprising a metadata search device that searches for stored content of a signal similar to a content signal and outputs metadata of the searched stored content, wherein the metadata management system includes: A metadata generation device, wherein the metadata search device includes information on a query interval indicating the content interval or region, and a size of the accumulated content interval or region of a signal similar to a signal included in the query interval. Section length correspondence acquisition unit for reading from section length correspondence storage unit that stores in advance section length correspondence information in association with storage section length information indicating the length, and information on a search section indicating a section or area of the content A search command acquisition unit for acquiring a signal similar to a signal included in the search interval; Based on the metadata search unit that extracts the metadata of the section or region of the tree, and the storage section length information of the section length correspondence information and the storage section length information corresponding to the query section included in the search section, A search result output unit that outputs the metadata extracted by the metadata search unit, and the metadata generation device includes the query interval and the accumulation interval of a signal similar to a content signal of the query interval; , A content correspondence acquisition unit that acquires a plurality of correspondence relationships, a joint correspondence generation unit that generates a joint correspondence by combining a plurality of correspondences between the query section and the storage section, and the joint correspondence Section length as the section length correspondence information including the query section combined by the generation unit and storage section length information indicating the size of the storage section of the connection correspondence A section length corresponding relation creation section for storing the response relation storage unit, a metadata search system comprising: a.

  According to the present invention, the metadata search device stores the metadata of the stored content section or region of the signal similar to the signal included in the content section or region, and indicates the stored content section or region size. Since the data is output based on the section length information, the metadata can be output according to the degree of similarity of the content, and the user or the like can quickly select the metadata that matches the content. For example, in the case of outputting metadata in descending order of stored content sections or areas, the metadata search device can output metadata in descending order of the degree of similarity of content, so that a user or the like can output the content. The matching metadata can be selected quickly. In addition, since the metadata search device outputs an appearance frequency histogram, it is possible to select metadata in a time interval with a high appearance frequency, and display metadata that matches the search content.

It is a conceptual diagram which shows the metadata search system which concerns on embodiment of this invention. It is a schematic block diagram which shows the structure of the metadata production | generation apparatus which concerns on this embodiment. It is a flowchart which shows an example of operation | movement of the metadata management apparatus which concerns on this embodiment. It is a conceptual diagram which shows an example of the coupling | bonding corresponding relationship which concerns on this embodiment. It is a conceptual diagram which shows an example of the content information which concerns on this embodiment. It is a conceptual diagram which shows an example of the some connection correspondence which concerns on this embodiment. It is a conceptual diagram which shows an example of the section length correspondence information which concerns on this embodiment. It is a schematic block diagram which shows the structure of the metadata search apparatus which concerns on this embodiment. It is a flowchart which shows an example of operation | movement of the metadata search apparatus which concerns on this embodiment. It is the schematic which shows an example of the query data search screen which concerns on this embodiment. It is the schematic which shows an example of the query data search screen which concerns on the modification 1 of this embodiment. It is the schematic which shows an example of the appearance frequency histogram which concerns on the modification 2 of this embodiment. It is a figure which shows another example of the metadata which the search result output part which concerns on the modification 2 of this embodiment displays. It is a figure which shows another example of the metadata which the search result output part which concerns on the modification 2 of this embodiment displays. It is a figure which shows another example of the metadata which the search result output part which concerns on the modification 2 of this embodiment displays.

(Embodiment)
Hereinafter, embodiments of the present invention will be described in detail with reference to the drawings. FIG. 1 is a conceptual diagram showing a metadata search system according to an embodiment of the present invention. In this figure, the metadata search system includes a metadata generation device a1, a database b1, and a metadata search device c1. The database b1 includes a correspondence relationship storage unit b111, a content information storage unit b121, a section length correspondence relationship storage unit b122, and an accumulated content storage unit b131. The content information storage unit b121 and the section length correspondence storage unit b122 are referred to as a metadata storage unit b12.

When new content is input, the metadata generation device a1 stores the input new content, the stored content of a signal similar to the signal of the content, and the stored content stored in the stored content storage unit b131 in advance. The combined correspondence information is generated and stored in the correspondence storage unit b111 (see FIG. 4). Here, the content input or stored in the metadata generation device a1 is, for example, a content such as a moving image or music, and the user uploads the metadata generation device a1 or automatically creates a website by a computer program. Content downloaded to the metadata generation device a1.
The metadata generation device a1 generates new content metadata based on the combined correspondence information stored in the correspondence storage unit b111 and the stored content metadata stored in the metadata storage unit b12. The metadata generation device a1 stores the generated metadata in the metadata storage unit b12. Note that the metadata is additional information for content, for example, content type, title, or information on performers / performers. Further, the metadata generation device a1 includes correspondence information between a new content section or area (query section) and a combined stored content section or area (storage section; storage section length information) in the metadata. Certain section length correspondence information is stored in the section length correspondence storage unit b122 (see FIG. 7). The metadata generation device a1 stores metadata (referred to as content information) other than the section length correspondence information in the content information storage unit b121 (see FIG. 5).

When a search command including a search condition designated by a user or the like is input, the metadata search device c1 outputs metadata of stored content that matches the search condition.
Here, the metadata search device c1 according to the present embodiment sorts and outputs the output order of the search result information based on the section length correspondence information stored in the section length correspondence storage unit b122.
Hereinafter, the metadata generation apparatus a1 will be described first, and then the metadata search apparatus c1 will be described.

<About Configuration of Metadata Generation Device a1>
FIG. 2 is a schematic block diagram showing the configuration of the metadata generation apparatus a1 according to this embodiment. In this figure, the metadata management apparatus a1 includes a content correspondence acquisition unit a101, a combined correspondence generation unit a102, a content information generation unit a103, and an accumulation section length correspondence generation unit a104. Further, the metadata management apparatus a1 receives the query content ID and the section or region of the query content, and the stored content ID and the section or region of the stored content, and outputs the query metadata to be added to the query data.

The content correspondence acquisition unit a101 acquires a plurality of correspondences including a query interval and a signal accumulation interval similar to the content signal in the query interval from the correspondence relationship storage unit b111 in FIG.
The combined correspondence generation unit a102 generates a combined correspondence by combining a plurality of correspondences between the query interval and the storage interval acquired by the content correspondence acquisition unit a101. Specifically, the combined correspondence generation unit a102 calculates a difference between time intervals (referred to as query time intervals) of query contents having a plurality of correspondence relationships and a difference between time intervals (referred to as accumulation time intervals) of the stored content. To do. The combined correspondence generation unit a102 combines the query time intervals and the accumulation time intervals of a plurality of correspondence relationships based on the calculated difference, the query content ID and the combined query time interval, and the stored content ID and the combined storage time. A connection correspondence consisting of sections is generated.

The content information generation unit a103 obtains the content information about the query section or the content information about the entire query content from the connection correspondence generated by the connection correspondence generation unit a102, and the stored content ID and metadata of the corresponding stored content. Based on at least one of the above. The content information generation unit a103 stores the generated content information in the content information storage unit b121 in FIG.
The section length correspondence generation unit a104 stores the query section combined by the combination correspondence generation unit a102 and the accumulated section length information as section length correspondence information in the section length correspondence storage unit b122 in FIG. To do.

<Operation of Metadata Management Device a1>
The processing from the content correspondence acquisition unit a101 to the content information generation unit a103 will be specifically described.
FIG. 3 is a flowchart showing an example of the operation of the metadata generation apparatus a1 according to this embodiment.

(Step S101) The content correspondence acquisition unit a101 reads the query content ID and query time interval of the query content, and the accumulated content ID and accumulation time interval of the accumulated content in the database corresponding to the query content. Then, a plurality of correspondence relationships between the query content ID and query time interval of the query content and the accumulated content ID and accumulation time interval of the accumulated content in the database corresponding to the query content are acquired.

  Here, the content correspondence acquisition unit a101 uses a search result obtained by searching for a corresponding section of sound or video by the method described in paragraph “0016” of JP-A-2008-219608, Get correspondence. Specifically, the content correspondence acquisition unit a101 is a feature sequence stored in advance by a storage unit (not shown), and a feature sequence (accumulation) of stored data divided into 16-second lengths every 3 seconds. (Referred to as feature series). Next, the content correspondence acquisition unit a101 applies a window (referred to as an input signal attention window) to the query content signal (referred to as an input signal), and generates a feature amount sequence for the signal within the input signal attention window (generation). The feature amount series obtained is called a query feature amount series). The content correspondence acquisition unit a101 calculates an input signal similarity value indicating the degree of similarity between the generated query feature quantity sequence and the read accumulated feature sequence. The content correspondence acquisition unit a101 calculates a movable amount indicating the amount by which the input signal attention window can be moved based on the calculated input signal similarity value, and based on the movable amount, the position of the input signal attention window And the input signal attention window is set at that position. Further, the content correspondence acquisition unit a101 calculates the input signal similarity value for each position of the input signal attention window by repeating these processes. The content correspondence acquisition unit a101 determines that the query feature value series and the stored feature series match (determine that they are similar) when the input signal similarity value is greater than a preset threshold value, and does not match when the input feature similarity value is equal to or less than the threshold value. judge.

If the content correspondence acquisition unit a101 determines that the query feature amount series and the stored feature series match, it generates a correspondence as follows. The content correspondence acquisition unit a101 extracts the query content file name used for the search as the query content ID, and extracts the input signal attention window section as the query time section. Then, the ID of the corresponding stored content obtained as the search output is extracted as the stored content ID, and the corresponding section output as a location corresponding to the query time section of the stored content is extracted as the storage time section. Note that association information of information (path, file name) indicating the storage location of the stored content ID (file name), the query time interval, and the stored feature series is stored in advance in the storage unit.
The content correspondence acquisition unit a101 generates a correspondence that associates the extracted query content ID, query time interval, accumulated content ID, and accumulation time interval. Thereafter, the process proceeds to step S102.

  Here, a time interval is used as the corresponding range, but a trajectory in a space combining time expansion and contraction of time and sound or pitch variation may be used as the interval, or a portion in the plane of the still image A region such as an image may be used, or a region combining a partial image in video and time may be used. Further, the correspondence relationship may be acquired by a method described in Japanese Patent Application Laid-Open No. 2006-524545 instead of the method described in Japanese Patent Application Laid-Open No. 2008-219608. In the present embodiment, the query content ID and the stored content ID are the file names of the query content and the stored content, but any information that identifies the query content and the stored content may be used.

(Step S102) The combined correspondence generation unit a102 selects one correspondence other than the correspondence selected in Step S102 up to the previous time, which is the correspondence acquired in Step S101 (the selected correspondence is selected). The first correspondence). Thereafter, the process proceeds to step S103.

(Step S103) The combined correspondence generation unit a102 obtains one correspondence relationship other than the correspondence relationship selected in Step S103 up to the previous step in the repetition of Steps S103 to S106, which is the correspondence relationship acquired in Step S101. Select (the selected correspondence is called the second correspondence). Thereafter, the process proceeds to step S104.
The correspondence relationship selected in step S103 is a correspondence relationship other than the first correspondence relationship selected in the immediately preceding step S102.
(Step S <b> 104) The combined correspondence generation unit a <b> 102 determines a combining condition based on the first corresponding relationship and the second corresponding relationship. When the joining condition is satisfied, the second correspondence is joined to the first correspondence (the joined correspondence is referred to as a joint correspondence). The generation process of this connection correspondence will be described later. Thereafter, the process proceeds to step S105.

(Step S105) The content information generation unit a103 generates content information corresponding to the query time interval from the combined correspondence generated in step S104. Thereafter, the process proceeds to step S106.
(Step S106) The section length correspondence generation unit a104 generates the query section and the storage section combined in step S104 as section length correspondence information. The generation process of this metadata (content information, section length correspondence information) will be described later. Thereafter, the process proceeds to step S107.

(Step S107) The combined correspondence generation unit a102 determines whether or not all the second correspondences have been selected, that is, whether or not the second correspondence has been processed. When it is determined that all the second correspondences have been processed (Yes), the process returns to step S103. On the other hand, if it is determined that all the second correspondences have not been processed (No), the process proceeds to step S108.
(Step S108) The combined correspondence generation unit a102 determines whether or not all the first correspondences have been selected, that is, whether or not the first correspondence has been processed. When it is determined that all the first correspondences have been processed (Yes), the process returns to step S102. On the other hand, when it is determined that all the first correspondence relationships have not been processed (No), the operation ends.
As described above, in the operation illustrated in FIG. 3, the combined correspondence generation unit a102 selects each of the first correspondence and the second correspondence from all the correspondences that have been read, and satisfies the condition. A connection correspondence is generated while determining whether or not to combine. Further, the section length correspondence generation unit a104 generates a query section and its accumulation section length information as section length correspondence information based on the generated combined correspondence.

<About the process for generating the association correspondence>
FIG. 4 is a conceptual diagram illustrating an example of the connection correspondence according to the present embodiment. The generation process of the connection correspondence is performed by the connection correspondence generation unit a102. This diagram is a diagram when music data having a file name “C2.mp3” is input as query content. In this figure, first, the first correspondence A (FIG. 4A) and the second correspondence B (FIG. 4B) are input, and the combined correspondence (FIG. 4C) is generated. It shows that. The generated combined correspondence is stored in the storage unit as a second correspondence (FIG. 4D). Next, the first correspondence relationship C (FIG. 4E) and the second correspondence relationship (FIG. 4D) are input, and the combined correspondence relationship (FIG. 4F) is generated. Show.

The combined correspondence generation unit a102 generates a combined correspondence relationship by combining the storage time intervals corresponding to the query content with respect to the correspondence relationship from the correspondence relationship A to the correspondence relationship C. Content information generating unit a103 is a can or can not bond begins to q ₁ of a first corresponding relationship query time interval, start the q ₂ of the second corresponding relationship query time interval, the d ₁ first the beginning of the accumulation time period, beginning the d ₂ of the second accumulation time interval, as a threshold value (5 seconds) that has been predetermined to theta, determines by using the following equation (1).

When the above expression (1) is satisfied, the combined correspondence generation unit a102 combines the first corresponding relationship and the second corresponding relationship. In other words, the expression (1) is the above coupling condition.
As the threshold value θ increases, the time interval of the combined correspondence relationship to be combined becomes longer, and the metadata corresponding to this time interval includes a lot of unrelated metadata. On the other hand, when the threshold value θ decreases, the time interval of the combined correspondence relationship to be combined becomes short, and related metadata cannot be combined. In the present embodiment, as a result of simulation, 5 seconds is selected as the threshold value θ that is an appropriate time interval in which relevant metadata is included. The threshold may be around 5 seconds to several seconds.

In the example of FIG. 4, the correspondence A shown in FIG. 4A is the first correspondence, the correspondence B shown in FIG. 4B is the second correspondence, and the left side of Expression (1) is “0”. Therefore, the expression (1) is satisfied. In this case, the combined correspondence generation unit a102 combines the correspondence A and the correspondence B as shown in FIG.
Since the combined correspondence shown in FIG. 4D is the first correspondence, and the correspondence C shown in FIG. 4E is the second correspondence, Equation (1) becomes “0”. Equation (1) is satisfied. In this case, the connection correspondence generation unit a102 combines the correspondence A and the connection correspondence between A and B as shown in FIG. As a result, as shown in FIG. 4F, the correspondence relationship A to the correspondence relationship C are combined to form one combined correspondence relationship. For example, FIG. 4F illustrates the query content ID “C2.mp3” having the query content “3:21” to “3:43”, and the stored content ID “C1.wav” having the stored content “14”. This indicates that the connection correspondence from “second” to “30 seconds” has been generated.
In the above example, it is determined whether or not combining is possible using the start of the time interval, but the end of the time interval may be used.

Further, the combined correspondence generation unit a102 may select the third correspondence relationship in addition to the first and second correspondence relationships, and use the ratio of the difference between the time intervals of these correspondence relationships as the combination condition. For example, binding relationship generation unit a102 is the binding conditions begin to q ₁ of a first correspondence between the query time interval, start the q ₂ of the second corresponding relationship query time interval, the q ₃ third Correspondence start of query time interval, d ₁ is start of first accumulation time interval, d ₂ is start of second accumulation time interval, d ₃ is start of third accumulation time interval, and θ is predetermined. As the set threshold value (0.5), the following equation (2) is used for determination. As the threshold value θ, 0.5 is selected as the threshold value θ that is an appropriate time interval in which relevant metadata is included as a result of simulation.

  When the above expression (2) is satisfied, the combined correspondence generation unit a102 combines the first corresponding relationship, the second corresponding relationship, and the third corresponding relationship. The combined correspondence generation unit a102 sets the smallest time among the beginnings of the time intervals of the correspondence relationship satisfying the combination condition as the start of the time interval after the combination, and sets the largest time among the end of the time intervals of the correspondence relationship. As the end of the time interval after the combination, this time interval is set as the time interval of the connection correspondence.

Moreover, although the case where the corresponding relationship was combined based on the time interval was described above, the present invention is not limited to this, and the corresponding relationship may be combined based on the position area. In this case, as the correspondence relationship, the content ID and the position area information are acquired for each area such as a partial image in the two-dimensional image. In this case, the connection correspondence generation unit a102 may determine whether to combine using the following connection conditions. For example, the combined correspondence generation unit a102 sets q ₁ = (q ₁₁ , q ₁₂ , q ₁₃ , q ₁₄ ) to the upper limit value, lower limit value, horizontal value of the vertical axis of the two-dimensional coordinates of the query content having the first correspondence relationship. A vector representing an area composed of the upper limit value and the lower limit value of the axis, and q ₂ = (q ₂₁ , q ₂₂ , q ₂₃ , q ₂₄ ) is set as the upper limit value, lower limit value of the vertical axis of the query content of the second correspondence relationship, A vector representing an area composed of an upper limit value and a lower limit value on the horizontal axis, and d ₁ = (d ₁₁ , d ₁₂ , d ₁₃ , d ₁₄ ) is set as an upper limit value, a lower limit value, A vector representing an area consisting of an upper limit value and a lower limit value on the horizontal axis, and d ₂ = (d ₂₁ , d ₂₂ , d ₂₃ , d ₂₄ ) is an upper limit value, a lower limit value on the upper left vertical axis of the second stored content, A vector representing the area consisting of the upper and lower limit values on the horizontal axis, and θ _{(Θ 1, θ 2, θ} 3, θ 4) as a threshold that has been predetermined to determine using the following equation (3) to (6).

  When all of the above formulas (3) to (6) are satisfied, the combined correspondence generation unit a102 combines the first corresponding relationship and the second corresponding relationship.

<About metadata generation processing>
The metadata generation process will be described below.
First, content information generation processing will be described.

  FIG. 5 is a conceptual diagram showing an example of content information according to the present embodiment. The content information generation unit a103 generates, as content information of new query content, content information obtained by adding content information of the corresponding stored content to the content information of the query content of the binding correspondence generated by the binding correspondence generation unit a102. To do. This figure shows the content information before and after the content information generation process performed based on the connection correspondence shown in FIG.

FIG. 5A is a schematic diagram showing content information before content information generation processing. This figure shows that the content information of the query content of the query time interval “03: 21-03: 40” of the query content ID “C2.mp3” is “Genre 1: Drama, Genre 2: Suspense, Performer: ΔΔHanako, broadcast medium: TV, BS, channel ID: 348Ch ”. Also, this figure shows that before the generation process, the content information of the stored content IDs “C1.wav” and “C2.mp3” are “music title: ABC, genre: rock, artist: Taro XX, broadcast medium: radio, AM, frequency: 1079 Hz ”. The broadcast medium is information indicating the type of broadcast medium such as a television or radio, BS (Broadcasting Satellite) broadcast, or terrestrial broadcast. The channel and frequency are information for identifying a broadcast frequency and information for identifying a broadcast station.
FIG. 5B is a schematic diagram showing the content information after the content information generation processing. This figure shows that after the generation process, the content information of the query content with the query content ID “C8.mpeg” is “genre 1: drama, genre 2: suspense, performer: △ Hanako, song name: ABC, genre: rock , Artist: Taro XX, Broadcast Medium: TV, BS, Channel ID: 348Ch ”. That is, the content information generation unit a303 generates content information obtained by adding the content information of the corresponding stored content (C1.wav) to the content information of the query content (C2.mpeg) as new content information of the query content. It shows that. It should be noted that information that is a predetermined item such as a broadcast medium, channel ID, or frequency and that is unique to the content may not be added as new content information. Further, the stored content ID may be included in the metadata. Moreover, although this embodiment demonstrated the case where the metadata to produce | generate was metadata of a query time area, this invention is not restricted to this, You may produce | generate metadata as metadata of the whole query content.
In this way, the metadata management apparatus a1 generates metadata obtained by adding the metadata of the stored content to the metadata of the query content as the metadata of the query content. Thereby, the metadata management apparatus a1 can add and enrich the metadata of the query content without requiring the user to add the metadata. For example, in the example of FIG. 5, music information is added to the query content of the video. In this way, the metadata management device can search for BGM (BackGround Music) music flowing in a video scene, for example, and add the searched music information to the metadata of the video query content.

Next, the section length correspondence information generation process will be described.
The section length correspondence generation unit a104 includes a query time section of the joint correspondence generated by the joint correspondence generation unit a102, and a storage time section (corresponding to a storage time section for each stored content that has been combined and converted into a query time section. Section length correspondence relationship information in which storage time sections; storage sections; storage section length information) are associated with each other is generated. 6 shows an example of the connection correspondence, and FIG. 7 shows an example of the section length correspondence information generated based on the connection correspondence shown in FIG.

  FIG. 6 is a conceptual diagram showing an example of a plurality of connection correspondences according to the present embodiment. This figure is an example of a connection correspondence when there are a plurality of stored contents corresponding to the query contents. In this figure, the stored content with the stored content IDs “C4.mpeg”, “C5.wav”, “C6.wmv”, “C7.mp3” corresponds to the query content with the query content ID “C3.mpeg”. The connection correspondence relationship is shown. Correspondences 1 to 4 are combined correspondences corresponding to the stored content IDs “C4.mpeg”, “C5.wav”, “C6.wmv”, and “C7.mp3”, respectively. For example, the connection correspondence 1 is changed from “0 minutes 0 seconds” to “3 minutes 50 seconds” of the query content ID “C3.mpeg” and “2 minutes” of the stored content ID “C4.mpeg”. “31 seconds” to “6 minutes 31 seconds” corresponds.

FIG. 7 is a conceptual diagram showing an example of section length correspondence information according to the present embodiment. This figure shows the section length correspondence information when the joint correspondence shown in FIG. 6 is input.
FIG. 7 illustrates the section length correspondence information of the query content of the time section “00: 00-03: 0” of the query content ID “C3.mpeg”. For example, this figure shows that the corresponding storage time interval of the storage content ID “C4.mpeg” of the signal similar to the signal of the query content “C3.mpeg” is “00: 00-03: 0”. The storage section length indicated by the corresponding storage time section is “180 seconds”. Further, for example, in this figure, the corresponding storage time intervals of the storage content ID “C5.wav” of the signal similar to the signal of the query content “C3.mpeg” are “00: 40-01: 30” and “02:10”. −03: 00 ”. The storage section length indicated by the corresponding storage time section is the total length of the corresponding storage time section “100 seconds”.

<Configuration of Metadata Search Device c1>
Hereinafter, the metadata search device c1 will be described.
FIG. 8 is a schematic block diagram showing the configuration of the metadata search apparatus c1 according to this embodiment. In this figure, the metadata search device c1 includes a section length correspondence acquisition unit c101, a search command acquisition unit c102, a metadata search unit c103, and a search result output unit c104. Also, the metadata search device c1 receives the query content ID of the section length correspondence information, the time section or region of the query content, the stored content ID, the section or region of the stored content, and the content information of the stored content. . Further, the metadata search device c1 outputs the search content ID and the search content section or region indicated by the search command, the corresponding stored content ID, the stored content section or region, and the content information.

The section length correspondence acquisition unit c101 reads the section length correspondence relation information from the section length correspondence storage unit b122.
The search command acquisition unit c102 acquires a time interval (referred to as a search interval) of content to be searched (referred to as search content).
The metadata search unit c103 extracts metadata (section length correspondence information, content information) of a stored content section or area of a signal similar to a signal included in the search section.
The search result output unit c104 arranges the metadata extracted by the metadata search unit c103 based on the storage section length information of the section length correspondence information and the storage section length information corresponding to the query section included in the search section. Change to output.

<Operation of Metadata Retrieval Device c1>
The processing from the section length correspondence acquisition unit c101 to the search result output unit c104 will be specifically described.
FIG. 9 is a flowchart showing an example of the operation of the metadata search apparatus c1 according to the present embodiment.

(Step S201) The section length correspondence acquisition unit c101 reads the section length correspondence information from the section length correspondence storage unit b122. Thereafter, the process proceeds to step S202.
(Step S202) The search command acquisition unit c102 waits for input of a search condition included in the search command. Thereafter, the process proceeds to step S202.
(Step S203) When a search command is input, the search command acquisition unit c102 reads a search condition included in the search command. Here, the search condition includes, for example, metadata and a search time interval. Thereafter, the process proceeds to step S204.
(Step S204) The metadata search unit c103 selects metadata corresponding to the search condition. Specifically, first, the metadata search unit c103 selects a search content ID from the search conditions read in step S203. For example, when a metadata keyword is input as a search condition, metadata including the keyword is selected from the metadata storage unit b12, and a file name corresponding to the metadata is selected as a search content ID. Next, the metadata search unit c103 selects the section length correspondence information in which the query content ID and the query time section are the search content ID and the search time section, respectively, from the section length correspondence information b122. Further, based on the stored content ID of the selected section length correspondence information, the content information of the stored content ID is extracted from the content information storage unit b121. Thereafter, the process proceeds to step S205.

(Step S205) The search result output unit c104 rearranges the metadata selected in step S204 in descending order of the accumulation section length. Specifically, the search result output unit c104 calculates the time (storage section length) included in the corresponding storage time section of the section length correspondence information, and rearranges the stored content IDs and storage sections in the order of the calculated time. For example, in the example of FIG. 7, the search result output unit c104 sets the storage section lengths of the storage content IDs “C4.mpeg”, “C5.wav”, “C6.wmv”, and “C7.mp3” to “ It is calculated as “180 seconds”, “100 seconds”, “80 seconds”, and “70 seconds”. In this case, the search result output unit c104 rearranges the stored content IDs in the order of the stored content IDs “C4.mpeg”, “C5.wav”, “C6.wmv”, “C7.mp3”. Thereafter, the process proceeds to step S206.
(Step S206) The search result output unit c104 outputs the storage content ID of the section length correspondence information, the storage section, and the content information of the storage content ID, that is, metadata in the order rearranged in Step S205. . The output information is displayed on a display unit such as a display in the output order, for example. Thereafter, the process proceeds to step S207.
(Step S207) The search result output unit c104 determines whether all search conditions have been processed. If it is determined that all search conditions have been processed (YES), the operation is terminated. If it is determined that all the search conditions have not been processed (NO), the process returns to step S202.

FIG. 10 is a schematic diagram illustrating an example of the query data search screen V1 according to the present embodiment. In this figure, the broadcast medium is “TV”, “terrestrial”, the channel ID is “019Ch.”, And the broadcast time zone (search time section) is “2009 April 11, 2009 09:00” to “2009. This shows a case where the search condition “April 11, 10:00” is input. In this case, the metadata search unit c103 selects, from the metadata storage unit b12, metadata including the broadcast medium “TV”, “terrestrial”, and the channel ID “019Ch.” As keywords, and corresponds to the metadata. The file name to be selected is selected as the search content ID. Next, the metadata search unit c103 has the query content ID as the search content ID and the query time interval from “April 11, 2009 09:00” to “April 11, 2009 10:00”. Certain section length correspondence information is selected from the section length correspondence information b122. Further, based on the stored content ID of the selected section length correspondence information, the content information of the stored content ID is extracted from the content information storage unit b121. That is, the metadata (interval length correspondence relationship information and content information) of the accumulated content including all the time intervals including the interval in which the correspondence relationship overlaps with the time interval designated as the query is acquired. The search result output unit c104 rearranges and displays the metadata selected and extracted by the metadata search unit c103 in descending order of accumulation section length.
FIG. 10 shows that metadata of 23 stored contents is extracted as a result of the above search conditions and processing, and a list of 8 of them is displayed.

Further, FIG. 10 shows that the metadata of the stored content of the search result is rearranged and displayed in descending order of the storage section length. For example, the metadata V11 of the stored content surrounded by a broken line with the reference symbol V11 has the broadcast medium “TV”, “BS”, the channel ID “348Ch.”, And the storage section “April 11, 2009”. "09:08 on the day" to "09:14 on April 11, 2009" and the accumulation section length is "7 minutes". In addition, the metadata V12 of the stored content surrounded by a broken line with the reference symbol V12 indicates that the broadcast medium is “radio”, “AM”, the frequency is “1079 Hz”, and the storage section is “09 April 03, 2009, 09:03”. From “Minute” to “April 11, 2009 09:05” and “April 11, 2009 09:45” to “April 11, 2009 09:46”, the accumulation section length is “5” "Minute".
The button displayed on the right side of the metadata indicates a radio button for selecting the metadata of the stored content. For example, in this figure, a radio button surrounded by a broken line with reference numeral V13 is selected, indicating that the stored content of the metadata V11 is selected. In this case, when the viewing button is pressed, the metadata search apparatus reads the stored content corresponding to the metadata from the stored content storage unit b131, plays it, and displays it on the screen.

  As described above, according to the present embodiment, the metadata search device c1 uses the storage section length information of the stored content as the metadata of the time section of the stored content of the signal similar to the signal included in the time section of the search content. Therefore, the metadata can be output according to the degree of similarity of the content, and the user or the like can quickly select the metadata that matches the certain content. Here, if the accumulation section length is long, the section where the signal is similar is also long, and it can be said that the degree of similarity of the content is high. Therefore, for example, when outputting metadata in the order of long accumulation interval length, the metadata search device c1 can output metadata in descending order of the degree of similarity of content, and the user or the like matches a certain content. Metadata can be selected quickly.

<Modification 1>
In the above embodiment, the search result output unit c104 outputs the corresponding accumulation time interval converted into the query time interval as the accumulation interval. However, the accumulation time interval may be output as it is, or the corresponding accumulation interval, the accumulation time interval, Both may be output.
FIG. 11 is a schematic diagram illustrating an example of a query data search screen V2 according to the first modification of the present embodiment. When the query data search screen V2 (FIG. 11) according to the present modification is compared with the query data search screen V1 (FIG. 10) according to the above embodiment, the items of metadata displayed as search results are different. FIG. 11 shows that an accumulation time interval is displayed as metadata. For example, the metadata V21 of the stored content surrounded by the broken line with the reference symbol V21 has the broadcast medium “TV”, “BS”, the channel ID “348Ch.”, And the storage time interval “April 2009”. 10th 13:32 "to" April 10, 2009 13:38 "and the accumulation section length is" 7 minutes ".
In general, there is a trend in video and music, and broadcasting or the like is performed at a certain time, and a user uploads contents to a server or the like at a certain time. According to the first modification, since the metadata search device c1 outputs the accumulation time interval, it is possible to select an accumulation time interval that is close to the search time interval, and to select metadata that matches the content. it can.

<Modification 2>
Further, the search result output unit c104 calculates a histogram representing the appearance frequency of the stored content for each section or region of the content in the search section, and the metadata search unit c103 extracts based on the appearance frequency indicated by the calculated histogram. The meta data may be output and displayed.
FIG. 12 is a schematic diagram illustrating an example of an appearance frequency histogram according to the second modification of the present embodiment. In this figure, the horizontal axis indicates time (search time interval), and the vertical axis indicates appearance frequency. In this figure, the time is divided every predetermined time (1 minute).
For example, the appearance frequency surrounded by the broken line H11 indicates the number of accumulated contents of signals similar to the search content signal between “9:10:00” and “9:10:59”, and the number Indicates “6”.

Here, the search result output unit c104 generates and outputs an appearance frequency histogram as follows. The search result output unit c104 first generates a histogram array and sets the appearance frequency in the histogram array to be empty (“0” times). Here, the histogram arrangement indicates, for example, the appearance frequency of each section obtained by dividing the search time section by a one-minute time section (bin width). The search result output unit c104 extracts an accumulation interval corresponding to the search time interval from the interval length correspondence information, and when the extracted time interval is included in the time interval indicated by the histogram array, the current frequency of the time interval is set to “1”. "Add times and increment.
The search result output unit c104 may output the appearance frequency indicated by the histogram attached to the metadata as the search result, or may search the metadata from the appearance frequency. Here, the search result output unit c104 may output information indicating a time section having a high appearance frequency indicated by the histogram as the search result, or the search results (for example, stored content ID and The content names) may be rearranged and output. Specifically, the search result output unit c104 sorts the stored contents in descending order of the peak value of the histogram. Alternatively, the histogram may be sorted in descending order of integral value.
That is, the search result output unit c104 determines the output order of search results based on the appearance frequency indicated by the acquired histogram. In this way, it becomes possible to prioritize the stored content using the information of the appearance frequency histogram.

  Content such as popular content that is highly preferred by users is frequently broadcast, and many users upload content to a server or the like. In other words, in a time interval in which the appearance frequency of the stored content is high, the search content is the stored content having a strong preference, and there is a high possibility that a plurality of stored content of the same signal exists. Therefore, in the time interval in which the appearance frequency is high, it is highly possible that the stored content that matches the search content can be selected and the metadata can be selected. According to the second modification, since the metadata search device c1 outputs the appearance frequency histogram, it is possible to select metadata in a time section with a high appearance frequency, and display metadata that matches the search content. it can.

13 to 15 are diagrams illustrating examples of metadata displayed by the search result output unit c104 according to the second modification of the present embodiment.
FIG. 13 is a screen displaying the path to the stored content and the file name (URL) of the stored content in descending order of the number of stored content IDs in the combined correspondence. Specifically, the search result output unit c104 extracts pairs of accumulated content IDs and query content IDs of all the connection correspondences, and eliminates duplication. Next, the search result output unit c104 aggregates the stored content IDs of the group from which duplication is eliminated, and displays the URLs of the stored content in descending order of the accumulated content IDs. Note that the search result output unit c104 counts the number of query contents including accumulated content by eliminating duplication, but the present invention is not limited to this, and duplication may not be eliminated. The fact that stored content is included in query content is called secondary use of stored content.
In this figure, for example, the stored content with the rank “1” is included in the query content with the secondary usage frequency “21”, and the URL is “http: //XXXXXX/database/cid0002931.flv”. It shows that.

FIG. 14 is a screen displayed when the URL indicated by reference numeral P1 in FIG. 13 is selected and clicked. FIG. 14 shows that the appearance frequency histogram of the content is displayed with the stored content selected in FIG. 13 as the query content.
For example, in FIG. 14, the time interval with the highest appearance frequency is the “9” time interval with the highest secondary usage frequency, and the time is “00: 00: 20-00: 0040”.
In this way, in FIG. 14, the search result output unit c104 outputs the appearance frequency histogram, so that it is possible to select the metadata in the time interval with a high appearance frequency, and to display the metadata that matches the search content. it can.

  FIG. 15 is a screen displayed when “Details” indicated by reference symbol P2 in FIG. 14 is selected and clicked. FIG. 15 shows the metadata of the stored content at the time (in this example, “00: 00: 20-00: 00: 40”) selected by the query content (content ID is “cid0002931”) in FIG. Indicates that For example, in the second metadata from the bottom of the list displayed at the top of the screen, the content is “video” media, and the time interval (referred to as the corresponding interval 1) on the query content included in the combined correspondence is “00:00”. : 00: 00-00: 23 ", the URL of the stored content is" http: //XXXXXX/database/cid6283258.flv ", and the section (referred to as the corresponding section 2) on the corresponding stored content is" 00: 00: 00- “00:00:23”.

Further, FIG. 15 shows that the metadata of the stored content is rearranged and displayed in descending order of the storage section length (the time section of the corresponding section 1). For example, the first and second metadata from the bottom of the list displayed at the top of the screen indicate that the storage section length is 22 seconds and 23 seconds, respectively.
In addition, when the first metadata from the top and the first metadata from the bottom are compared, the corresponding section 2 regardless of the same media (“music”) and the same URL (accumulated content ID “cid02424456.flv”). Is different. In general, in music, the same melody and lyrics may be repeated (the first time is called the first chorus, the second time is called the second chorus). That is, the first metadata from the top is cid02424456. The entire music signal in flv is extracted as a similar signal, and the first metadata from the bottom is cid02424456. It shows that a part of the second chorus signal of the music in flv is extracted as a signal similar to the first chorus signal of the query content (cid0002931.flv).
The image at the bottom of the screen is a moving image when the corresponding section “00: 00: 00-00: 00: 23” indicated by P3 is clicked. The search target video G1 indicates that the video of the corresponding section 1 is displayed in the query content (content ID is “cid0002931”). Further, the matching video G2 indicates that a video at a time corresponding to the time displayed as the search target video G1 is displayed in the stored content (content ID “cid 6283258.flv”). FIG. 15 shows that the matching video G2 is a video in which a character string (telop) is added to the search target video G1. In this way, the user can obtain additional information by displaying the stored content corresponding to the query content.

In the above-described embodiment, the metadata search device c1 includes the configuration of the metadata generation device a1 (content correspondence acquisition unit a101, combined correspondence generation unit a102, content information generation unit a103, and storage section length correspondence generation unit a104. ).
In the embodiment, the metadata search device c1 may output only the metadata of the stored content having the longest storage section length. Thereby, the metadata search device c1 can select and output the metadata of the stored content having the highest degree of content similarity. Further, the metadata search device c1 may output metadata of a predetermined number of stored contents in order of increasing storage section length.
In the above embodiment, the metadata search device c1 may output metadata based on the appearance frequency histogram. For example, only the metadata of the time interval with the highest appearance frequency may be output.

  Note that a part of the metadata generation device a1 and the metadata search device c1 in the above-described embodiment, for example, the content correspondence acquisition unit a101, the combined correspondence generation unit a102, the content information generation unit a103, and the storage section length correspondence generation The unit a104, the section length correspondence acquisition unit c101, the search command acquisition unit c102, the metadata search unit c103, and the search result output unit c104 may be realized by a computer. In that case, the program for realizing the control function may be recorded on a computer-readable recording medium, and the program recorded on the recording medium may be read by a computer system and executed. Here, the “computer system” is a computer system built in the metadata generation apparatus a1 or the metadata search apparatus c1, and includes an OS and hardware such as peripheral devices. The “computer-readable recording medium” refers to a storage device such as a flexible medium, a magneto-optical disk, a portable medium such as a ROM and a CD-ROM, and a hard disk incorporated in a computer system. Furthermore, the “computer-readable recording medium” is a medium that dynamically holds a program for a short time, such as a communication line when transmitting a program via a network such as the Internet or a communication line such as a telephone line, In such a case, a volatile memory inside a computer system serving as a server or a client may be included and a program that holds a program for a certain period of time. The program may be a program for realizing a part of the functions described above, and may be a program capable of realizing the functions described above in combination with a program already recorded in a computer system.

  As described above, the embodiment of the present invention has been described in detail with reference to the drawings. However, the specific configuration is not limited to the above, and various design changes and the like can be made without departing from the scope of the present invention. It is possible to

  a1 ... metadata generation device, b1 ... database, c1 ... metadata search device, b11 ... correspondence storage unit, b121 ... content information storage unit, b122 ... interval length correspondence Storage unit, b131 ... accumulated content storage unit, b12 ... metadata storage unit, a101 ... content correspondence acquisition unit, a102 ... combined correspondence generation unit, a103 ... content information generation unit, a104: Accumulation section length correspondence generation unit, c101 ... Section length correspondence acquisition unit, c102 ... Search command acquisition unit, c103 ... Metadata search unit, c104 ... Search result output unit

Claims (11)

In a metadata search device for searching for stored content of a signal similar to a content signal and outputting metadata of the searched stored content,
A search result for outputting metadata of the stored content section or region of a signal similar to a signal included in the content section or region based on storage section length information indicating the size of the stored content section or region A metadata search apparatus comprising an output unit. In a metadata search device that searches for stored content of a signal similar to the content signal and outputs metadata of the searched stored content,
A section length in which information on a query section indicating the section or area of the content is associated with storage section length information indicating a size of the section or area of the stored content of a signal similar to a signal included in the query section A section length correspondence acquisition unit that reads from a section length correspondence storage unit that stores correspondence information in advance;
A search command acquisition unit for acquiring information of a search section indicating the section or area of the content;
A metadata search unit that extracts metadata of a section or region of the stored content of a signal similar to a signal included in the search section;
A search result output unit that outputs the metadata extracted by the metadata search unit based on the storage section length information corresponding to the query section included in the search section, which is the storage section length information of the section length correspondence information When,
A metadata search apparatus comprising:   The search result output unit stores the meta data in the descending order of the section or area size indicated by the storage section length information corresponding to the query section included in the search section, which is the storage section length information of the section length correspondence information. 3. The metadata search apparatus according to claim 2, wherein the metadata extracted by the data search unit is output.   The metadata search according to claim 2 or 3, wherein the storage section length information is storage section information indicating a section or a region of the stored content of a signal similar to a signal included in the query section. apparatus.   5. The search result output unit outputs accumulation section length information corresponding to a query section included in the search section, which is storage section length information of the section length correspondence information. The metadata search device according to any one of the items.   The search result output unit calculates a histogram representing the appearance frequency of the stored content for each section or region of the content in the search section, and the metadata search unit extracts based on the appearance frequency indicated by the calculated histogram The metadata search apparatus according to any one of claims 2 to 5, wherein the metadata is output.   The metadata search apparatus according to claim 6, wherein the search result output unit determines an output order of metadata extracted by the metadata search unit based on an appearance frequency indicated by the calculated histogram. A content correspondence acquisition unit that acquires a plurality of correspondences including the query interval and the accumulation interval of a signal similar to the content signal of the query interval;
A combined correspondence generation unit that generates a combined correspondence by combining a plurality of correspondences between the query section and the storage section;
The query section combined by the combined correspondence generation unit and the storage section length information indicating the size of the storage section of the combined correspondence are stored in the section length correspondence storage unit as the section length correspondence information. A section length correspondence generation unit;
The metadata search device according to any one of claims 2 to 7, further comprising: In a metadata search method in a metadata search apparatus for searching for stored content of a signal similar to a content signal and outputting metadata of the searched stored content,
The section length correspondence acquisition unit includes information on a query section indicating the section or area of the content, and storage section length information indicating the size of the stored content section or area of a signal similar to a signal included in the query section. And a section length correspondence acquisition process that reads from a section length correspondence storage unit that prestores section length correspondence information associated with each other,
A search command acquisition unit acquires a search section information indicating a section or area of the content;
A metadata search process for extracting metadata of a section or region of the stored content of a signal similar to a signal included in the search section;
The metadata extracted by the search result output unit in the metadata search process based on the storage section length information of the section length correspondence information and the storage section length information corresponding to the query section included in the search section A search result output process for outputting
A metadata search method characterized by comprising: In the computer of the metadata search device that searches the stored content of the signal similar to the signal of the content and outputs the metadata of the searched stored content,
A section length in which information on a query section indicating the section or area of the content is associated with storage section length information indicating a size of the section or area of the stored content of a signal similar to a signal included in the query section Section length correspondence acquisition means for reading from the section length correspondence storage unit that stores correspondence information in advance,
Search command acquisition means for acquiring information on a search section indicating the section or area of the content,
Metadata search means for extracting metadata of a section or region of the stored content of a signal similar to a signal included in the search section;
Search result output for outputting metadata extracted by the metadata search means based on storage section length information corresponding to a query section included in the search section, which is storage section length information of the section length correspondence information means,
Metadata search program to function as. In a metadata management system including a metadata search device that searches for stored content of a signal similar to a content signal and outputs metadata of the searched stored content,
The metadata management system comprises a metadata generation device,
The metadata search device includes:
A section length in which information on a query section indicating the section or area of the content is associated with storage section length information indicating a size of the section or area of the stored content of a signal similar to a signal included in the query section A section length correspondence acquisition unit that reads from a section length correspondence storage unit that stores correspondence information in advance;
A search command acquisition unit for acquiring information of a search section indicating the section or area of the content;
A metadata search unit that extracts metadata of a section or region of the stored content of a signal similar to a signal included in the search section;
A search result output unit that outputs the metadata extracted by the metadata search unit based on the storage section length information corresponding to the query section included in the search section, which is the storage section length information of the section length correspondence information When,
With
The metadata generation device includes:
A content correspondence acquisition unit that acquires a plurality of correspondences including the query interval and the accumulation interval of a signal similar to the content signal of the query interval;
A combined correspondence generation unit that generates a combined correspondence by combining a plurality of correspondences between the query section and the storage section;
The query section combined by the combined correspondence generation unit and the storage section length information indicating the size of the storage section of the combined correspondence are stored in the section length correspondence storage unit as the section length correspondence information. A section length correspondence generation unit;
A metadata search system comprising:

Images