JP2007164633A - Content retrieval method, system thereof, and program thereof - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To display result obtained by retrieving specified contents and arranging meta information related to the matter of the content. <P>SOLUTION: When character information is inputted as a search query, contents having related meta information with high degree of similarity to the query are selected, and sameness of the selected contents are judged based on the feature quantity of the extracted contents, and the contents whose sameness are high are grouped and displayed. When a content is inputted as a retrieval query, the feature quantity of the content of the query is extracted, the contents having the feature quantity whose sameness is high with the feature quantity of the content of the query are selected, and these degrees of similarity are determined based on the related meta information which the contents have, grouped and displayed. <P>COPYRIGHT: (C)2007,JPO&INPIT

Description

  The present invention relates to a content search method, apparatus, and program, and in particular, selects and presents content that is highly identical to the content used as a search query for a set of digital content such as moving images, still images, audio, and text information. The present invention relates to a content search method, apparatus, and program.

  Information processing devices such as personal computers and mobile terminals, optical fibers, 3G mobile communications, etc. have become widespread, and various contents such as moving images, still images, audio, text information are distributed and used in many homes and businesses. . Under such an environment, there is a need for a search method in which a user can efficiently find desired content from a large amount of information. Also, in the distribution process of content, the provider may be distributed under unintended usage conditions. The provider can confirm the location of the content, and the user can search for content distributed under valid usage conditions. There is a need for a content search method for this purpose.

  The content search method includes a method in which a user inputs meta information related to content as a search query, and a method in which the feature amount of the content itself is used. In the former method, a robot or the like visits the Web and automatically collects meta information related to the content, and performs text matching based on the collected metadata to identify the desired content. In the latter case, there is a method of extracting feature quantities specific to the content based on features such as the color, shape, and pitch of the content and identifying them by matching those feature quantities (see Non-Patent Document 1, for example) ).

In addition, there is also a method in which an approximate ranking of a search query and other content is calculated from both the above-described content feature amount and meta information, and a search result is output based on the ranking (see, for example, Patent Document 1).
"High-speed image retrieval system based on objects: ExSight", Kazuhiko Kushima et al., IPSJ Transactions, Vol.40, No.2, pp732-741, (1999). JP-A-11-224256, “Information Retrieval Method and Recording Medium Recording Information Retrieval Program”

  In the above-described conventional content search, when content-related meta information is used in a search query, the meta information does not always accurately represent the content, and the content that is completely different from the desired content is the search result. It is displayed as, and the work to narrow down while confirming the result is a heavy burden.

  On the other hand, if the content feature amount is used in the search query, content with similar contents can be obtained as a search result, but it is not easy to collect related meta information of the displayed content and measure the reliability of the search result. . For example, it is assumed that an original high-quality image is searched for a certain thumbnail image obtained from the Web connected to the Internet, and is desired to be used together with related meta information. At this time, according to the conventional search method, the content and related meta information are obtained as unstructured information, the necessary information is found from various different meta information, and the reliability of the information in view of the obtained meta information as a whole. It was difficult to estimate sex. For example, even in the conventional method disclosed in Patent Document 1, since meta information is presented as accompanying information, the displayed search result is valid. It cannot be easily determined.

  The present invention has been made in view of the above points, and searches for specific contents, organizes them from meta information related to the contents of the contents, displays the results, and allows the user to easily obtain desired contents and related meta data. It is an object of the present invention to provide a content search method, apparatus, and program capable of obtaining information and assisting confirmation of reliability.

  FIG. 1 is a diagram for explaining the principle of the present invention.

The present invention (Claim 1) is a content search method for searching for specific content from a content set and presenting it.
When the meta information collecting means acquires the related meta information of the content as the search query from the user, the meta information collecting step (step 1) for collecting the related meta information of the content corresponding to the search query from the content set;
The meta information identification means compares the search query with the related meta information of the collected content to determine the relevance of the information, and from the content DB that stores the related meta information and feature quantity of the content in association with each other. A meta-information identification step (step 2) for acquiring relevant content;
A content feature amount acquisition means for extracting a content feature amount for each content from the content DB (step 3);
A group classification step (group classification unit) determines the identity of the content based on the content feature amount obtained in the content feature amount acquisition step, and groups the content determined to have an identity higher than a predetermined value ( Step 4) and
The output means performs an output step (step 5) for outputting at least one of the content and the related meta information of the content for each group classified in the group classification step.

Further, the present invention (Claim 2) is an output step (Step 5) of Claim 1,
In order from the group including the content determined to be highly relevant in the group classification step (step 4), at least one of the content and the related meta information of the content is output for each group for a certain group or all groups. To do.

Further, the present invention (Claim 3) is as described in Claim 1 or 2,
A content registration step of extracting a feature amount of the designated content in advance and associating the designated content and the extracted feature amount of the content with the related meta information of the input content and registering them in the content registration DB; ,
A registered meta information acquisition step of searching the content registration DB using the feature quantity of the content having the identity in the group classification step (step 4) as a search query and acquiring the registered related meta information;
In the output step (step 5),
At the same time as outputting at least one of the content and the related meta information of the content for each group classified in the group classification step (step 4), the registered related meta information acquired in the registered meta information acquisition step is output.

The present invention (Claim 4) is a content search method for searching for specific content from a content set and presenting it.
A search query acquisition step in which a search query acquisition means acquires content as a search query from a user;
A query content feature amount acquisition unit extracts a feature amount A of content input as a search query, and a query content feature amount extraction step;
The feature amount comparison unit extracts the feature amount B of the content acquired from the content DB in which the related meta information and the feature amount of the content are associated and accumulated, compares the feature amount A with the feature amount B, and A feature amount comparison step of determining identity and acquiring content determined to be higher than a predetermined value;
A meta information acquisition unit for collecting related meta information of the content corresponding to the content acquired in the feature amount comparison step;
A group classification step in which the group classification means determines the relevance of the content based on the related meta information obtained in the meta information acquisition step, and groups the content determined to have high relevance;
An output means performs an output step of outputting at least one of content and related meta information of the content for each group classified in the group classification step.

Further, the present invention (Claim 5) provides an output step according to Claim 4,
At least one of the content and the related meta information of the content is output for each group with respect to a certain number of groups or all the groups in order from the group including the content determined to have high identity by the group classification step.

Further, the present invention (Claim 6) is as described in Claim 4 or 5,
A content registration step of extracting a feature amount of the designated content in advance and associating the designated content and the extracted feature amount of the content with the related meta information of the input content and registering them in the content registration DB; ,
The registration meta-information acquisition step of searching the content registration DB using the feature amount of the content that became the same in the group classification step as a search query and acquiring the registered related meta-information is further performed.
In the output step,
At least one of the content and the related meta information of the content is output for each group classified in the group classification step, and at the same time, the registered related meta information acquired in the registered meta information acquisition step is output.

Further, the present invention (Claim 7) is the group classification step of Claim 4,
After selecting content with a certain identity from the content set, the content is grouped according to identity,
In the output step,
For each group, at least one of content and related meta information of the content is output.

Further, the present invention (Claim 8) is the output step of Claim 7,
In order from the group of contents determined to have high identity by the group classification step, at least one of the content and the related meta information of the content is output for each group for a certain number of groups or all groups.

Further, according to the present invention (Claim 9), the feature amount of the specified content is extracted in advance in Claim 7 or 8, and the specified content and the extracted feature amount of the content are input. A content registration step of associating the related meta information of the content and registering it in the content registration DB;
In the group classification step, the registration meta information acquisition step of searching the content registration DB using the feature amount of the content having the identity as a search query and acquiring the registered related meta information is further performed.
In the output step,
At least one of the content and the related meta information of the content is output for each group classified in the group classification step, and at the same time, the registered related meta information acquired in the registered meta information acquisition step is output.

Further, the present invention (Claim 10) according to Claim 3, 6 or 9,
License information in the related meta information of each content accumulated in the content registration DB by the content registration step (hereinafter referred to as registered license information), and license information in the related information meta information collected in the registration meta information acquisition step (hereinafter referred to as “registered license information”). , The license license determination step for determining the license conditions, and the collected license information)
In the group classification step,
The inclusion relationship is judged by the license permission judgment step and classified into groups.
In the output step,
At least one of content and related meta information of the content is output for each group.

Further, the present invention (claim 11) is the method according to claim 10,
A step of determining an inclusion relationship in the license permission determination step;
In the output step,
At least one of the content associated with the collected license information within the permitted range of the registered license information and the related meta information of the content is output.

  FIG. 2 is a principle configuration diagram of the present invention.

The present invention (Claim 12) is a content search device for searching for specific content from a content set and presenting it.
A content DB 160 that stores content-related meta information and feature quantities in association with each other;
Meta information collection means 110 that collects related meta information of content corresponding to a search query from a content set when content related meta information is acquired as a search query from a user;
Meta information identifying means 120 that compares the search query with the related meta information of the collected content, determines the relevance of the information, and acquires content having a certain relevance from the content DB 160;
Content feature amount acquisition means 130 for extracting content feature amounts for each content from the content DB 160;
A group classification unit 140 that determines the identity of the content based on the content feature amount obtained by the content feature amount acquisition unit 130, and groups the content determined to have an identity higher than a predetermined value;
Output means 150 for outputting at least one of the content and the related meta information of the content for each group classified by the group classification means 140.

Further, the present invention (Claim 13), in Claim 12,
Content registration means for extracting a feature amount of designated content in advance and registering the designated content and the extracted feature amount of the content in association with the related meta information of the content in the content registration DB; ,
A registration meta-information acquiring unit that searches the content registration DB using the feature amount of the content having the identity by the group classification unit as a search query, and acquires the related meta-information that is registered;
The output means is
At least one of the content and the related meta information of the content is output for each group classified by the group classification unit, and at the same time, the registered related meta information acquired by the registered meta information acquisition unit is output.

The present invention (Claim 14) is a content search device for searching for specific content from a content set and presenting it.
A content DB that stores content-related meta information and feature quantities in association with each other;
Search acquisition means for acquiring content as a search query from a user;
Query content feature amount extraction means for extracting feature amount A of the content input as a search query;
The feature amount B of the content acquired from the content DB is extracted, the feature amount A and the feature amount B are compared, the identity of the content is determined, and the content whose identity is determined to be higher than a predetermined value is acquired Characteristic amount comparison means
Meta information acquisition means for collecting related meta information of the content corresponding to the content acquired by the feature amount comparison means;
Group classification means for determining the relevance of the content based on the related meta information obtained by the meta information acquisition means, and grouping the content determined to be highly relevant;
Output unit for outputting at least one of content and related meta information of the content for each group classified by the group classification unit.

Further, the present invention (Claim 15) according to Claim 14,
Content registration means for extracting a feature amount of designated content in advance and registering the designated content and the extracted feature amount of the content in association with the related meta information of the content in the content registration DB; ,
A registered meta-information acquiring unit that searches the content registration DB using the feature amount of the content that is the same by the group classification unit as a search query and acquires the registered related meta-information;
The output means is
At least one of the content and the related meta information of the content is output for each group classified by the group classification unit, and at the same time, the registered related meta information acquired by the registered meta information acquisition unit is output.

The present invention (Claim 16) is characterized in that in Claim 14,
Content registration means for extracting a feature amount of designated content in advance and registering the designated content and the extracted feature amount of the content in association with the related meta information of the content in the content registration DB; ,
A registration meta-information acquisition unit that searches the content registration DB using the feature quantity of the content having the identity in the group classification unit as a search query and acquires the registered related meta-information;
The output means is
At least one of the content and the related meta information of the content is output for each group classified by the group classification unit, and at the same time, the registered related meta information acquired by the registered meta information acquisition unit is output.

Further, the present invention (Claim 17) according to Claim 13, 15, or 16,
License information in the related meta information of each content accumulated in the content registration DB by the content registration means (hereinafter referred to as registered license information) and license information in the related information meta information collected in the registered meta information acquisition step (hereinafter referred to as “registered license information”). And license license determination means for determining a license condition,
Group classification means
The inclusion relationship is judged by the license permission judgment means and classified into groups,
The output means is
At least one of content and related meta information of the content is output for each group.

  The present invention (Claim 18) is a content search program for causing a computer to function as the content search apparatus according to Claims 12 to 17.

  As described above, according to the present invention, it is possible to search for specific contents and receive meta information related to the contents of contents in an organized manner.

  Furthermore, the user can easily obtain the content having the necessary license conditions according to the purpose of use.

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

[First Embodiment]
In the present embodiment, a content search when character information (related meta information) is input from a user as a search query will be described.

  FIG. 3 is a configuration diagram of the content search apparatus according to the first embodiment of the present invention.

  An input device 100 for inputting a search query, a network 20 connected to the value Web server 30, and a display device 40 for displaying search results are connected to the content search device 100 shown in FIG.

  The search device 100 includes a meta information collection unit 110, a meta information identification unit 120, a content feature amount acquisition unit 130, a group classification unit 140, an output unit 150, a content DB 160, and a classification information DB 170. The content DB 160 is a database that stores related meta information and feature amounts of content in association with each other.

  The operation in the above configuration will be described.

  FIG. 4 is a flowchart of the operation of the content search apparatus according to the first embodiment of the present invention.

  Step 101) The content search device 100 receives a search query (meta information related to content (hereinafter referred to as related meta information) from the input device 10.

  Step 102) The meta information collection unit 110 collects related meta information of content in the content set based on the input search query. For example, if the content set is an image published on a Web site (Web server 30) connected via the network 20, a robot that circulates the Web site, collects HTML files, and analyzes text information is used. It is also possible to obtain the tag and the character information designated by the tag in the HTML file including the image information or the HTML file of the site linked from the HTML file.

  Step 103) Next, the meta information identifying unit 120 compares the related meta information collected by the meta information collecting unit 110 with the search query to determine the relevance. The content-related meta information includes, for example, a title, an author name, an author's group, a keyword, license information, and the like.

  Step 104) Then, all contents determined to be highly relevant (for example, having a similarity equal to or higher than a set threshold) are selected and stored in a memory (not shown) in the meta information identification unit 120.

  Step 105) The content feature amount acquisition unit 130 acquires the feature amount of the content stored in the content DB 160 based on the content selected by the meta information identification unit 120, and the memory in the content feature amount acquisition unit 130 (Not shown).

  Step 106) Next, the group classification unit 140 acquires content feature amounts from a memory (not shown) in the content feature amount acquisition unit 130, and refers to the classification information DB 170 to determine mutual identity. The contents determined to be the same are classified into the same group and stored in a memory (not shown) in the group classification unit 140. The determination of the identity based on the feature amount is performed based on, for example, a character string partial matching ratio stored in the classification information DB 170, a difference value as numerical information, a partial matching ratio in a bit string, and the like.

  Step 107) The output unit 150 presents the selected content in the group classified state on the display device 40 as a search result from the memory (not shown) of the group classification unit 140. The presentation method may be the content itself, or one or all of the related metadata. There are cases where all the contents in the group are displayed, and there is a method of displaying one representative content in the group.

  Further, the order of the group of contents to be displayed can be set in the order in which the relevance of the meta information is high. FIG. 5A shows an example in which four image contents are selected from a certain content group using the keyword “family car” as a search query. Assume that related meta information “family car”, “family sedan”, “sports car”, and “popular car” are attached in the order in which the relevance is determined to be high. As shown in FIG. 5B, the group classification unit 140 groups two contents having a feature quantity “0001”, and the output unit 150 includes meta information contents most relevant to the search query. Displayed in order from the group (the group with the feature amount “0001”).

  On the other hand, it is also possible to display in descending order of the average value of the relevance rank of the groups grouped by the feature amount. For example, as shown in FIG. 5C, content having the feature amount “FFF1” is displayed. There is also a method of making the first group.

  Further, in the process of collecting the meta information in step 102 described above, the meta information obtained by the collection activity using the network 20 is accumulated in advance in a DB (meta information DB), and the DB It is also possible to obtain meta information by accessing.

[Second Embodiment]
FIG. 6 shows the configuration of the search device according to the second embodiment of the present invention. In this figure, the same parts as those in FIG.

  In this embodiment, a content feature amount extraction unit 210, a registration unit 220, a content registration DB 230, and a registration meta information acquisition unit 240 are added to the first embodiment described above.

  FIG. 7 is a flowchart of the operation of the content search apparatus in the second embodiment of the present invention.

  Since the processing from step 101 to step 106 is the same as that in the first embodiment, the description thereof will be omitted.

  The following steps 201 to 204 are performed before the steps 101 to 106 are performed.

  Steps 201 and 202) The content is input to the content feature amount extraction unit 210 and the content related meta information (hereinafter referred to as registration meta information) from the input device 10 in advance to the registration unit 220.

  Step 203) The content feature amount extraction unit 210 extracts feature amounts from the input content in the same manner as the content feature amount acquisition unit 130 described above.

  Step 204) The registration unit 220 stores the content feature amount extracted by the content feature amount extraction unit 210 and the input registration meta information in the content registration DB 230 as shown in FIG. The content data itself may or may not be stored in the content registration DB 230.

  Step 205) In step 106, the registration meta information acquisition unit 240 searches the content registration DB 230 using the content features of each group classified by the group classification unit 140 as a search query, and registers the content determined to be the same. Meta information is acquired and output to the output unit 150 as a search result.

  Step 206) The output unit 150 presents the selected content on the display device 40 in a state where the search results are classified into groups. At the time of presentation, the obtained registration meta information is displayed for each group.

  Further, in the present embodiment, by registering a license describing the use permission condition in the content registration DB 230, the license is acquired and compared with the meta information licenses collected in the processes of steps 201 to 204. Determine the license inclusion relationship. The inclusion relationship is whether one condition matches the other condition or satisfies the other condition. For example, the content registration DB 230 has a permission condition “available until January 1, 2010”. When a license is acquired, the usage period is the same, or before or after. The group is further divided into small groups according to the license inclusion relationship.

  When presenting the search results, the output unit 150 presents them together for each small group, and displays the obtained license inclusion relationship. In addition, among the small groups, only those whose license inclusion relationship satisfies the license of the content registration DB 230 can be displayed. For example, in the above example, only the content associated with the license whose usage period is set before January 1, 2010 is displayed as the search result.

[Third Embodiment]
In the present embodiment, a case where content is input as a search query from a user will be described.

  FIG. 9 shows the configuration of a content search apparatus according to the third embodiment of the present invention. In this figure, the same components as those in FIGS.

  The configuration shown in the figure includes a search query acquisition unit 310, a query content feature amount acquisition unit 320, a feature amount comparison unit 330, a meta information acquisition unit 340, a group classification unit 140, an output unit 150, a content DB 160, a meta information DB 350, a classification. It consists of information DB170.

  Note that the meta information DB 350 is unnecessary when collecting meta information via a network.

  FIG. 10 is a flowchart of the operation of the content search apparatus in the tenth embodiment of the present invention.

  Step 301) A user specifies and inputs content from the input device 10.

  Step 302) The search query acquisition unit 310 extracts the feature amount A from the input content, and generates a search query.

  Step 303) The query content feature quantity acquisition unit 320 acquires a content set in the content DB 160, acquires a feature quantity B of the content from the content set, and a memory (not shown) in the query content feature quantity acquisition unit 320. Stored).

  Step 304) The feature quantity comparison unit 330 compares the feature quantity B stored in the memory (not shown) in the content feature quantity acquisition unit 320 with the feature quantity A of the search query to determine identity. To do.

  Step 305) As a result, all contents determined to have high identity are selected and stored in a memory (not shown) in the feature amount comparison unit 330.

  Step 306) The meta information acquisition unit 340 acquires content from a memory (not shown) in the feature amount comparison unit 330, and acquires related meta information related to the content from the meta information DB 350 or the network.

  Step 307) The group classification unit 140 refers to the classification information DB 170, compares some or all of the related meta information, classifies the contents determined to be highly related, and classifies the classification results into the same group. Store in a memory (not shown).

  Step 308) The output unit 150 obtains from the memory (not shown) in the group classification unit 140, and presents the selected content on the display device 40 in a state of being classified into groups.

  The presentation method may be the content itself or one or all of the related metadata. Further, there are cases where all the contents in the group are displayed, and there is a method of displaying one representative content in the group.

  In the process of extracting the content feature amount in step 303, the feature amount extraction operation is executed in advance for part or all of the content in the content set so that the feature amount is accumulated in the content DB 160, and the content DB 160 is accessed. By doing so, a method of making it possible to refer to the feature amount is also possible.

  Further, in the present embodiment, the order of the group of contents to be displayed can be an order including contents with high relevance of meta information. FIG. 11A is an example in which four image contents are selected from a certain content group using a content having a feature quantity [0001] as a search query. It is assumed that the feature quantity has a higher identity as the numerical value is closer, and associated meta information “family car”, “popular car”, “sport car”, and “family car” is attached in descending order of the identity. As shown in FIG. 11B, two contents whose meta information is “family car” are grouped, and a group including a content with a feature amount having the highest identity with the search query (group of meta information “family car”). ) Are displayed in order. On the other hand, it is also possible to display in the descending order of the average value of the feature quantity identity ranks of groups grouped by meta information. For example, as shown in FIG. 11C, the meta information is “popular car”. There is also a method in which the content is set to the first group.

[Fourth Embodiment]
FIG. 12 shows the configuration of a content search apparatus according to the fourth embodiment of the present invention. In this figure, the same components as those in FIGS. 3, 6, and 9 are denoted by the same reference numerals, and description thereof is omitted.

  The content search apparatus illustrated in FIG. 12 includes a search query acquisition unit 310, a query content feature amount acquisition unit 320, a feature amount comparison unit 330, a meta information acquisition unit 340, a group classification unit 140, a registered meta information acquisition unit 410, and an output unit 150. , Content DB 160, meta information DB 350, classification information DB 170, content feature amount extraction unit 210, registration unit 220, and content registration DB 230.

  FIG. 13 is a flowchart of the operation of the content search apparatus in the fourth embodiment of the present invention.

  In the registration flow, as in the second embodiment, when a specific content and its content-related meta information are input in advance (steps 201 and 202), feature amounts are extracted for the content (step 203). The content feature amount and the content related meta information are stored in the content registration DB 230 (step 204).

  The search flow is the same as that in the third embodiment until grouping is performed using the feature amounts in steps 301 to 307.

  After the above processing, the registration meta information acquisition unit 410 acquires the feature quantity of the group content stored in the memory (not shown) of the group classification unit 140, and uses the content registration DB 230 as a search query. Search is performed, and the registered meta information of the content that is determined to have high identity (for example, the difference in the feature amount is within the set threshold) is acquired and output to the output unit 150.

  The output unit 150 presents the selected content on the display device 40 as a search result in a state of being classified into groups. When presenting, the registered meta information of the obtained content is displayed for each group.

  In the present embodiment, by registering a license in which content usage permission conditions are registered in the content registration DB 230, the license information is acquired and collected in the registration flow (steps 201 to 204). Compared with the license, the registered meta information acquisition unit 410 further divides the group acquired from the group classification unit 140 into smaller groups based on the license inclusion relationship.

  When presenting the search result in the output unit 150, the small group is presented together on the display device 40, and the obtained license inclusion relationship is displayed.

  In addition, it is possible to display only small groups satisfying the license whose license inclusion relationship is stored in the content registration DB 230.

[Fifth Embodiment]
FIG. 14 shows the configuration of a content search apparatus according to the fifth embodiment of the present invention. In the figure, the same components as those in FIG. 9 are denoted by the same reference numerals, and the description thereof is omitted.

  The configuration shown in the figure includes a search query acquisition unit 310, a query content feature amount acquisition unit 320, a feature amount comparison unit 330, a feature amount group classification unit 510, a meta information acquisition unit 520, an output unit 150, a meta information DB 350, and a content DB 160. Consists of

  FIG. 15 is a flowchart of the operation of the content search apparatus in the fifth embodiment of the present invention.

  In the present embodiment, the processing in steps 301 to 305 is the same as the processing in FIG. 10 of the third embodiment described above.

In step 305, with respect to the content selected by the feature amount comparison unit 330, the feature amount group classification unit 510 classifies the content according to the identity based on the feature amount, and stores a memory (not shown) together with the feature amount. (Step 501).

  Next, the meta information acquisition unit 520 acquires each related meta information via the meta information DB 350 or the network based on the content of the group and associates the related meta information with the group, together with the feature amount, in the meta information acquisition unit 520 Is stored in a memory (not shown) (step 502). In the output unit 150, the selected information is presented to the display device 40 in the grouped state by using the meta information classified into the group stored in the memory (not shown) of the meta information acquiring unit 520 as a search result. (Step 503).

  The presentation method may be the content itself or one or all of the meta information related to the content. There are cases where all the contents in the group are displayed, and there is a method of displaying one representative content in the group.

  In the present embodiment, the order of groups of contents to be displayed can be an order including contents with high feature quantity identity. FIG. 16A shows an example in which four image contents are selected from a certain content group using a content having a feature quantity [0001] as a search query. It is assumed that the feature quantity has the higher identity as the numerical value is closer, and associated meta information “family car”, “popular car”, “sport car”, and “family car” is attached in descending order. As shown in FIG. 16B, two contents having a feature quantity “0001” are grouped and displayed in order from a group of contents having a feature quantity having the highest identity with the search query.

  According to the present embodiment, it is possible to collectively check what kind of meta information is attached to a content group having high content identity, and from the degree of attachment of statistical meta information for a certain content, It is possible to infer reliability.

[Sixth Embodiment]
FIG. 17 shows the configuration of a content search apparatus according to the sixth embodiment of the present invention.

  In this figure, the same components as those in FIGS. 14, 6 and 12 are designated by the same reference numerals, and the description thereof is omitted.

  The configuration illustrated in FIG. 17 is a configuration in which a content feature amount extraction unit 210, a registration unit 220, a content registration DB 230, and a registered meta information acquisition unit 610 are added to the configuration illustrated in FIG. 14 of the fifth embodiment.

  FIG. 18 is a flowchart of the operation of the content search apparatus in the sixth embodiment of the present invention.

  The registration flow to the content registration DB 230 is the same as that in steps 201 to 204 in FIG. 13 of the fourth embodiment. The content with high feature quantity identity is selected, and the meta information of the selected content is changed to meta information. The processing from step 301 to step 502 acquired from the DB 350 is the same as in the fifth embodiment.

  Thereafter, the registered meta information acquisition unit 610 acquires the feature amount of each group of content stored in the memory (not shown) of the meta information acquisition unit 520, and uses the feature registration DB 230 as a search query. The registration meta information of the content determined to be the same is acquired and stored in a memory (not shown) in the registration meta information acquisition unit 610 (step 601). The output unit 150 acquires registered meta information as a search result from a memory (not shown) in the registered meta information acquisition unit 610, and presents the selected content on the display device 40 in a state of being classified into groups ( Step 602). At the time of presentation, the obtained registration meta information is displayed for each group.

  In the present embodiment, by registering a license in which content usage permission conditions are registered in the content registration DB 230, the license is acquired and compared with the content meta information license collected in steps 201 to 204 of the registration flow. The group is further divided into small groups according to the license inclusion relationship. When presenting the search result, it is presented together for each small group, and the inclusion relation of the obtained license is displayed. In addition, it is possible to display only the small groups whose license inclusion relation satisfies the license of the content registration DB 230.

  In the present invention, the operations in the first to sixth embodiments described above are constructed as a program, installed in a computer used as a content search device, executed, or distributed via a network. Is possible.

  Further, only the operation of the registration flow (steps 201 to 204) in the second, fourth, and sixth embodiments may be an independent program as a content registration program.

  Further, the constructed program can be stored in a portable storage medium such as a hard disk, a flexible disk, or a CD-ROM, and installed in a computer to be executed or distributed.

  The present invention is not limited to the above-described embodiment, and various modifications and applications can be made within the scope of the claims.

  The present invention is applicable to a technique for searching for highly identical content from a large amount of information.

It is a figure for demonstrating the principle of this invention. It is a principle block diagram of this invention. It is a block diagram of the content search apparatus in the 1st Embodiment of this invention. It is a flowchart of operation | movement of the content search device in the 1st Embodiment of this invention. It is a figure which shows the example of a display in the 1st Embodiment of this invention. It is a block diagram of the content search device in the 2nd Embodiment of this invention. It is a flowchart of operation | movement of the content search device in the 2nd Embodiment of this invention. It is an example of content registration DB in the 2nd Embodiment of this invention. It is a block diagram of the content search device in the 3rd Embodiment of this invention. It is a flowchart of operation | movement of the content search device in the 3rd Embodiment of this invention. It is a display example in the 3rd Embodiment of this invention. It is a block diagram of the content search device in the 4th Embodiment of this invention. It is a flowchart of operation | movement of the content search device in the 4th Embodiment of this invention. It is a block diagram of the content search device in the 5th Embodiment of this invention. It is a flowchart of operation | movement of the content search apparatus in the 5th Embodiment of this invention. It is a display example in the 5th Embodiment of this invention. It is a block diagram of the content search device in the 6th Embodiment of this invention. It is a flowchart of operation | movement of the content search device in the 6th Embodiment of this invention.

Explanation of symbols

DESCRIPTION OF SYMBOLS 10 Input apparatus 20 Network 30 Web server 40 Display apparatus 100 Search apparatus 110 Meta information collection means, meta information collection part 120 Meta information identification means, meta information identification part 130 Content feature-value acquisition means, Content feature-value acquisition part 140 Group classification means , Group classification unit 150 output means, output unit 160 content DB (database)
170 Classification information DB
210 content feature amount extraction unit 220 registration unit 230 content registration DB (database)
310 search query acquisition unit 320 query content feature amount acquisition unit 330 feature amount comparison unit 340 meta information acquisition unit 350 meta information DB (database)
410 registered meta information acquisition unit 510 feature amount group classification unit 520 meta information acquisition unit 610 registered meta information acquisition unit

Claims (18)

A content search method for searching and presenting specific content from a content set,
When the meta information collecting means acquires the related meta information of the content as the search query from the user, the meta information collecting step of collecting the related meta information of the content corresponding to the search query from the content set;
The meta information identification unit compares the search query with the collected related meta information of the content, determines the relevance of the information, and determines from the content DB in which the related meta information of the content and the feature amount are stored in association with each other. A meta-information identifying step for acquiring content having a relevance of
A content feature amount acquiring unit that extracts a content feature amount for each content from the content DB;
Group classification in which group classification means determines the identity of the content based on the content feature amount obtained in the content feature amount acquisition step, and groups the content determined to have an identity higher than a predetermined value Steps,
An output step for outputting at least one of content and content-related meta information for each group classified in the group classification step;
The content search method characterized by performing. In the output step,
In order from the group including the content determined to be highly relevant by the group classification step, for at least one of the content and the related meta information of the content for each group for a certain group or all groups,
The content search method according to claim 1. A content registration step of extracting a feature amount of the designated content in advance and associating the designated content and the extracted feature amount of the content with the related meta information of the input content and registering them in the content registration DB; ,
The registration meta-information acquisition step of searching the content registration DB using the feature quantity of the content having the identity by the group classification step as a search query and acquiring the registered related meta-information is further performed.
In the output step,
Outputting at least one of content and content-related meta information for each group classified in the group classification step, and simultaneously outputting the registered related meta-information acquired in the registration meta-information acquisition step;
The content search method according to claim 1 or 2. A content search method for searching and presenting specific content from a content set,
A search query acquisition step in which a search query acquisition means acquires content as a search query from a user;
A query content feature amount acquiring unit for extracting a feature amount A of the content input as the search query;
The feature amount comparison unit extracts the feature amount B of the content acquired from the content DB in which the related meta information and the feature amount of the content are associated and stored, compares the feature amount A with the feature amount B, and the same content A feature amount comparison step for determining a gender and acquiring content determined to have a higher identity than a predetermined value;
A meta information acquisition step in which meta information acquisition means collects related meta information of the content corresponding to the acquired content;
A group classification step for determining the relevance of the content based on the related meta information obtained in the meta information acquisition step, and grouping the content determined to have high relevance;
An output step of outputting at least one of content and related meta information of the content for each of the groups classified in the group classification step;
The content search method characterized by performing. In the output step,
Outputting at least one of the content and the related meta-information of the content for each group for a certain number of groups in order from the group including the content determined to have high identity by the group classification step, or all groups;
The content search method according to claim 4. A content registration step of extracting a feature amount of the designated content in advance and associating the designated content and the extracted feature amount of the content with the related meta information of the input content and registering them in the content registration DB; ,
Further, a registration meta-information acquisition step of searching the content registration DB using the feature amount of the content made identical in the group classification step as a search query and acquiring the registered related meta-information is performed.
In the output step,
Outputting at least one of content and content-related meta information for each group classified in the group classification step, and simultaneously outputting the registered related meta-information acquired in the registration meta-information acquisition step;
The content search method according to claim 4 or 5. In the group classification step,
After selecting content with a certain identity from the content set, the content is grouped according to identity,
In the output step,
For each group, output at least one of content and related meta information of the content.
5. The content search method according to claim 4. In the output step,
In order from the group of contents determined to have high identity by the group classification step, for at least one of the content and the related meta-information of the content for each group for a certain number of groups or all groups,
The content search method according to claim 7. In advance, the feature quantity of the designated content is extracted from the contents in the same group by the content identity determination, and the designated content and the extracted feature quantity of the content are input. A content registration step of registering the related meta information in association with the content registration DB;
In the group classification step, a registration meta information acquisition step of searching the content registration DB using a feature amount of content having the same as a search query and acquiring the registered related meta information is performed.
In the output step,
Outputting at least one of content and content-related meta information for each group classified in the group classification step, and simultaneously outputting the registered related meta-information acquired in the registration meta-information acquisition step;
The content search method according to claim 7 or 8. License information (hereinafter referred to as registration license information) in the related meta information of each content accumulated in the content registration DB by the content registration step, and the related information meta information collected by the registration meta information acquisition step Further performing a license permission determination step of determining permission conditions for the license information (hereinafter referred to as collected license information);
In the group classification step,
The inclusion relationship is determined by the license permission determination step and classified into groups,
In the output step,
Outputting at least one of content and related meta information of the content for each group;
The content search method according to claim 3, 6 or 9. A step of determining an inclusion relationship in the license permission determination step;
In the output step,
The content search method according to claim 10, wherein at least one of content associated with collected license information within a licensed range of the registered license information and related meta information of the content is output. A content search device that searches for specific content from a content set and presents the content.
A content DB that stores content-related meta information and feature quantities in association with each other;
Meta information collection means for collecting related meta information of content as a search query from a user, and collecting related meta information of content corresponding to the search query from a content set;
Meta information identifying means for comparing the search query and related meta information of the collected content to determine the relevance of the information and acquiring content having a certain relevance from the content DB;
Content feature amount acquisition means for extracting a content feature amount for each content from the content DB;
Group classification means for determining the identity of the content based on the content feature quantity obtained by the content feature quantity acquisition means, and grouping the contents determined to have an identity higher than a predetermined value;
Output means for outputting at least one of content and content-related meta-information for each group classified by the group classification means;
A content search apparatus comprising: Content registration means for extracting a feature amount of designated content in advance and registering the designated content and the extracted feature amount of the content in association with the related meta information of the content in the content registration DB; ,
A registration meta-information acquisition unit that searches the content registration DB using the feature quantity of the content having the identity by the group classification unit as a search query and acquires the registered related meta-information;
The output means includes
Outputting at least one of the content and the related meta information of the content for each group classified by the group classification unit, and simultaneously outputting the registered related meta information acquired by the registered meta information acquisition unit;
The content search apparatus according to claim 12. A content search device that searches for specific content from a content set and presents the content.
A content DB that stores content-related meta information and feature quantities in association with each other;
Search acquisition means for acquiring content as a search query from a user;
Query content feature amount extraction means for extracting the feature amount A of the content input as the search query;
Content feature B obtained from the content DB is extracted, the feature A is compared with the feature B, content identity is determined, and content whose identity is determined to be higher than a predetermined value A feature amount comparison means for obtaining
Meta information acquisition means for collecting related meta information of the content corresponding to the content acquired by the feature amount comparison means;
Group classification means for determining the relevance of content based on the related meta information obtained by the meta information acquisition means, and grouping the content determined to be highly relevant;
Output means for outputting at least one of content and related meta information of the content for each group classified by the group classification means;
A content search apparatus comprising: Content registration means for extracting a feature amount of designated content in advance and registering the designated content and the extracted feature amount of the content in association with the related meta information of the content in the content registration DB; ,
A registration meta-information acquisition unit that searches the content registration DB using the feature amount of the content that is the same by the group classification unit as a search query, and acquires the registered related meta-information;
The output means includes
Outputting at least one of the content and the related meta information of the content for each group classified by the group classification unit, and simultaneously outputting the registered related meta information acquired by the registered meta information acquisition unit;
The content search device according to claim 14. Content registration means for extracting a feature amount of designated content in advance and registering the designated content and the extracted feature amount of the content in association with the related meta information of the content in the content registration DB; ,
The group classification means further includes registered meta information acquisition means for searching the content registration DB using a characteristic amount of content having identity as a search query and acquiring registered related meta information,
The output means includes
Outputting at least one of the content and the related meta information of the content for each group classified by the group classification unit, and simultaneously outputting the registered related meta information acquired by the registered meta information acquisition unit;
The content search device according to claim 14. License information (hereinafter referred to as registration license information) in the related meta information of each content accumulated in the content registration DB by the content registration means, and the related information meta information collected in the registration meta information acquisition step License license determination means for determining license conditions for license information (hereinafter referred to as collected license information),
The group classification means includes:
The inclusion relationship is judged by the license permission judging means and classified into groups,
The output means includes
Outputting at least one of content and related meta information of the content for each group;
The content search device according to claim 13, 15 or 16. Computer
18. A content search program that functions as the content search device according to claim 12.

Images