JP5226241B2 - How to add tags - Google Patents

Description

  The present invention relates to a method for adding a tag to moving image data, a server, and a program.

  Conventionally, a keyword called a tag is often associated with content on the Internet in order to facilitate the search for the content. Unlike an index that is automatically assigned by the management server and is not visible to the user, the tag registers a keyword that characterizes the content of the user. As a result, the user can easily grasp the contents. Furthermore, provision of effective content can be expected by referring to the search.

Here, content on the Internet includes various types of data, and in recent years, there are Web sites that manage moving image data. In such a situation, a technique for creating a search index for moving image data has been provided. For example, Patent Document 1 discloses an apparatus that extracts keywords from audio of video (moving image data) and automatically assigns an index to the video.
JP 2002-171481 A

  According to the above-described technique, an index for searching for a scene in the moving image data is created based on the sound included in the moving image data. This is simply extracted from a keyword included in the moving image data. Therefore, the moving image is not necessarily characterized. Therefore, it is not suitable for a search for selecting a target one from a plurality of moving image data including other moving images.

  Therefore, the present invention automatically adds a tag for characterizing a certain moving image data among a plurality of other moving image data, and the search engine effectively uses the moving image data or a Web page including the moving image data. It is an object to provide a method that enables extraction.

  For the above purpose, the following are specifically provided.

(1) A method of attaching a tag to moving image data,
Converting audio included in the moving image data into text data;
Extracting keywords from the text data;
Calculating the importance of the keyword in a predetermined search system;
Adding the keyword as a tag of the moving image data when the calculated importance satisfies a predetermined condition.

  According to such a configuration, the server that executes the method converts the voice included in the moving image data into text data, extracts the keyword from the text data, and calculates the importance of the keyword in a predetermined search system. When the calculated importance satisfies a predetermined condition, a keyword is assigned as a tag of moving image data.

  Thus, the server can assign, as a tag, a keyword whose importance in the search system satisfies a predetermined condition among keywords included in the audio of the moving image data, that is, a keyword that characterizes the moving image data. As a result, an effective tag based on information in consideration of other contents obtained from the search system can be automatically assigned.

  (2) The method according to (1), wherein the importance is based on the number of search results by the keyword in the predetermined search system.

  According to such a configuration, since the server executing the method uses the number of search results (number of hits) based on keywords as the importance in the search system, keywords appearing on more Web pages are moving image data. Selected as the tag for.

  As a result, a keyword familiar to the user is tagged, so that there is a possibility that the moving image data or a Web page including the moving image data is effectively searched.

  (3) The method according to (1), wherein the degree of importance is based on the number of times the search by the keyword is executed in the predetermined search system.

  According to such a configuration, the server that executes the method uses the number of times the search by the keyword is executed (search keyword ranking) as the importance in the search system. Is selected as a tag of moving image data.

  As a result, keywords frequently input by the user are preferentially tagged, so that there is a possibility that the moving image data or a Web page including the moving image data is effectively searched.

  (4) The degree of importance is further based on TF · IDF, which is a product of TF (Term Frequency) and IDF (Inverse Document Frequency), which is an index related to the appearance frequency of the keyword (2) or (3 ) Method.

  According to such a configuration, the server executing the method uses an index based on the appearance frequency of the keyword TF / IDF in addition to the number of hits and the search keyword ranking as the importance in the search system.

  Thus, for example, even if the number of hits or search keyword ranking does not satisfy a predetermined condition, if the keyword is characterized by the TF / IDF, the keyword is selected as a tag. Can do.

  As a result, the possibility of missing an important keyword can be reduced based on a single index, and it can be extracted as a tag characterizing a moving image.

  (5) The method according to any one of (1) to (4), wherein the keyword is a combination of a plurality of words.

  According to such a configuration, the server that executes the method performs tagging by determining the importance of combinations of a plurality of words. This makes it possible to express the content more accurately than the case where the assigned tag is a single word. Therefore, by selecting based on the number of hits, search keyword ranking, etc. in the search system, There is a possibility that an effective tag can be given.

  (6) When the keyword has a predetermined degree of similarity or more with category data stored in advance, the method further includes a step of assigning the category data as a tag of the moving image data. The method of crab.

  According to such a configuration, the server that executes the method can assign the category data as a tag by determining the similarity with the category data prepared in advance. Thus, for example, when category data is included in the sound of moving image data, the moving image data can be automatically classified into categories.

(7) Regarding the keyword, determining a field based on a predetermined rule;
The method according to any one of (1) to (6), further including a step of assigning data indicating the determined field as a tag of the moving image data.

  According to such a configuration, a server that executes the method determines a field based on a predetermined rule with respect to a keyword, and assigns data indicating the determined field as a tag of moving image data.

  Thus, the server can classify the moving image data based on the ontology (such as dictionary data used for classification) prepared in advance based on a predetermined rule (field estimation technique). As a result, based on the audio information of the moving image data, the field can be automatically estimated and a corresponding tag can be assigned.

(8) A server for attaching a tag to moving image data,
Means for converting audio contained in the moving image data into text data;
Means for extracting a keyword from the text data;
Means for calculating the importance of the keyword in a predetermined search system;
A server provided with the keyword as a tag of the moving image data when the calculated importance satisfies a predetermined condition.

  According to such a configuration, the same effect as in (1) can be expected by operating the server.

(9) A program for attaching a tag to moving image data,
Converting audio included in the moving image data into text data;
Extracting keywords from the text data;
Calculating the importance of the keyword in a predetermined search system;
A program for causing a server to execute the step of assigning the keyword as a tag of the moving image data when the calculated importance satisfies a predetermined condition.

  According to such a configuration, the same effect as in (1) can be expected by executing the program on the server.

  According to the present invention, a tag for characterizing certain moving image data among a plurality of other moving image data is automatically assigned, and the moving image data or a Web page including the moving image data is retrieved by a search engine. It can be extracted effectively.

  An example of a preferred embodiment according to the present invention will be described below based on the drawings.

[System overall configuration]
FIG. 1 is a conceptual diagram showing the overall configuration of a content management system according to an example of a preferred embodiment of the present invention.

  The tag determination server 10, the search server 20, the content server 30, and the terminal device 40 are connected via a communication line. In the content server 30, the content transmitted (posted) by the user of the terminal device 40 is managed and stored in the content DB.

  Here, the content posted by the user includes moving image data. The content server 30 requests the determination of the tag for the moving image data by transmitting the moving image data to the tag determination server 10.

  The tag determination server 10 requests search information from the search server 20 in order to acquire information to be referred to when determining a tag. The search server 20 executes search based on the keyword received from the tag determination server 10 or extracts data collected in advance, and provides the tag determination server 10 with the data. Thereby, the tag determination server 10 determines a tag for the moving image data by a process described later, and transmits the tag to the content server 30.

  The tag determined in this way can be displayed in response to a request from the user by storing it in the content DB in association with moving image data. Further, since this tag is referred to when searching for moving image data or content including moving image data, a search process based on the contents of moving image data is possible.

  The content management system in the present embodiment is realized by a plurality of servers, but the configuration is not limited to this, and the tag determination server 10 may have the functions of the search server 20 and the content server 30 together. Moreover, you may divide the tag determination server 10 into a some server according to each function.

[Function configuration]
FIG. 2 is a block diagram showing each function of the tag determination server 10 according to an example of the preferred embodiment of the present invention.

  The tag determination server 10 includes a voice recognition unit 110, a morpheme analysis unit 120, a tag candidate extraction unit 130, a search information collection unit 140, and a tag determination unit 150, and outputs tag data in response to input of moving image data. .

  The voice recognition unit 110 extracts voice data included in the received moving image data and converts it into text data using an existing voice recognition technique. As a result, information included in the moving image data is extracted as character information.

  The morpheme analysis unit 120 performs morpheme analysis on the text data generated by the speech recognition unit 110 and divides it into words. Here, an existing technique can be used for the morphological analysis method.

  The tag candidate extraction unit 130 extracts words (keywords) that can be assigned as moving image data tags from the words divided by the morphological analysis unit 120. Specifically, it is desirable to determine extraction rules in advance, such as limiting the part of speech to nouns and verbs.

  The search information collection unit 140 makes an inquiry to the search server 20 as reference information for determining a tag to be actually assigned from the tag candidates extracted by the tag candidate extraction unit 130, and collects the search information. (Details will be described later).

  The tag determining unit 150 refers to the search information acquired by the search information collecting unit 140, extracts a high-value search keyword from tag candidates, and determines it as a tag for moving image data. In this way, value collection on the Web can be performed by collecting information from the search server 20, so that more effective search keywords can be tagged than when tagging with information obtained only from moving image data. Can be attached.

  The tag determined in this way is stored in the content server 30 in association with moving image data. Specifically, for example, as in the tag data table shown in FIG. 4, it is desirable to store a plurality of tag data for each moving image data. Although tag data is given to a moving image, it may be given to a Web page including the moving image.

[Processing flow]
FIG. 3 is a diagram showing a flow of tag determination processing in the tag determination server 10 according to an example of the preferred embodiment of the present invention.

  In step S <b> 101, the tag determination server 10 receives moving image data to be tagged from the content server 30 and delivers it to the voice recognition unit 110.

  In step S102, the voice recognition unit 110 extracts voice data from the moving image data received in step S101. The moving image data often includes audio information, and by extracting this, information indicating the contents of the moving image data can be acquired.

  In step S103, the voice recognition unit 110 converts the voice data extracted in step S102 into text data. Specifically, it can be realized using existing technology, and subsequent language processing can be performed by using character information (text data).

  In step S104, the morphological analysis unit 120 performs morphological analysis on the text data generated in step S103, and decomposes it into a plurality of words. Thereby, a plurality of words are obtained as data indicating the contents of the moving image data.

  In step S105, the tag candidate extraction unit 130 extracts a tag candidate (keyword) from among the plurality of words acquired in step S104. Specifically, for example, tag candidates can be narrowed down by extracting limited parts of speech such as nouns and verbs.

  Further, the keyword may be a combination of a plurality of words. For example, by extracting a combination of nouns and verbs, etc., the content can be specified more, so that it may be suitable as a tag. In that case, it is desirable to perform extraction in consideration of the relationship between words, for example, using a dependency analysis technique.

  In step S106, the search information collection unit 140 collects search information from the search server 20 based on the tag candidate keywords extracted in step S105. Specifically, for example, the search server 20 executes a Web search based on the keyword received from the tag determination server 10. As a result, for example, the search number table shown in FIG. 5 is obtained.

  The number of search results based on keywords (the number of hits) is stored in the search number table. Here, for example, “XX restaurant” indicates that “1234 hits” were found as a search result.

  Further, the search server 20 can also manage the keyword input frequency as statistical information, as in the keyword-specific input count table shown in FIG. 6, for example, and provide the data of this table to the tag determination server 10. Good.

  In the keyword-specific number-of-inputs table, the number of times of input for each keyword for the search executed in the search server 20 is stored together with the date. As a result, the search server 20 can obtain the rank of the keyword input frequency within a predetermined period, and may provide this rank to the tag determination server 10.

  Further, the search server 20 may obtain and store this rank at a predetermined cycle and timing. For example, the keyword ranking table of FIG. 7 stores the ranking (ranking) of the number of keyword inputs for each year.

  Note that the ranking of the input frequencies may be performed by the tag determination server 10, and in that case, the search information collection unit 140 receives the data of the keyword-specific input count table and calculates the rank.

  In step S107, the tag determination unit 150 determines the importance of the keyword based on the search information acquired in step S106. Specifically, for example, the number of hits in the search number table (FIG. 5) is 1000 or more, the number of times of input in the previous month in the keyword-specific input number table (FIG. 6) is 10,000 or more, keyword ranking A determination condition is set in advance such that the ranking of the previous month in the table (FIG. 7) is within 10,000.

  For example, when using the search number table in FIG. 5 based on such a determination condition, “XX restaurant” is “1234 items” and the number of hits is 1000 or more, so it is selected as a tag.

  Further, in the case of using the keyword-specific input count table of FIG. 6, “XX restaurant” is selected as a tag because the input count “12345” of “April 2007” is 10,000 or more. The aggregation period may be in units of one month as described above, or may be a total value of preset periods such as the past six months.

  In the case of using the keyword / ranking table of FIG. 7, “XX restaurant” is “4567” and “ZZZ hotel” is “888” and is within 10000, so it is selected as a tag. Become.

  Here, when the number of hits and the ranking do not satisfy these conditions, for example, the importance level of the keyword may be determined by TF / IDF. In this case, if the TF / IDF value of the keyword is equal to or greater than a threshold value stored in advance, it can be determined that the keyword characterizes the moving image data and has high importance.

  The TF value related to the appearance frequency in the document (the moving image data) may be calculated each time. However, the IDF value related to the appearance frequency in a large number of document sets is calculated in advance and stored in the search server 20. Can be kept.

  In step S108, since it is determined that the importance of the keyword is equal to or higher than the predetermined value, the tag determination unit 150 assigns the keyword as a tag to the moving image data. Specifically, the keyword is transmitted to the content server 30 and stored in the tag data table (FIG. 4).

  In step S109, since the importance of the keyword is not greater than or equal to a predetermined value, the tag determining unit 150 determines that the keyword is not suitable as a tag, and determines a category as a process for assigning another tag.

  Specifically, for example, as shown in FIG. 8, a category matching or similar to the keyword is determined by similarity determination (matching) with previously stored category data. Further, category determination is not limited to such similarity determination. For example, as shown in FIG. 9, field data can be determined by using field estimation technology that refers to a predetermined ontology (such as dictionary data).

  In step S110, the tag determination unit 150 adds the category data or field data determined in step S109 to the moving image data as a tag. Specifically, as in step S108, category data or field data is transmitted to the content server 30 and stored in the tag data table (FIG. 4).

  The category and field are determined in accordance with the keyword importance determination in step S107, but steps S109 to S110 may be always performed regardless of the importance determination. In this case, both the keyword extracted from the moving image data and the data indicating the category and field are assigned as tags.

[Hardware configuration of server]
FIG. 10 is a diagram illustrating an example of a hardware configuration of the tag determination server 10 described in FIG. The tag determination server 10 includes a CPU (Central Processing Unit) 1010 (a plurality of CPUs such as a CPU 1012 may be added in a multiprocessor configuration), a bus line 1005, a communication I / F 1040, a main memory 1050, A BIOS (Basic Input Output System) 1060, a USB port 1090, an I / O controller 1070, a keyboard and a mouse 1100, and other input means and a display device 1022 are provided.

  Storage means such as a tape drive 1072, a hard disk 1074, an optical disk drive 1076, and a semiconductor memory 1078 can be connected to the I / O controller 1070.

  The BIOS 1060 stores a boot program executed by the CPU 1010 when the tag determination server 10 is started up, a program depending on the hardware of the tag determination server 10, and the like.

  The hard disk 1074 constituting the storage unit 107 stores various programs for the tag determination server 10 to function as a server and programs for executing the functions of the present invention, and various databases can be configured as necessary. .

  As the optical disc drive 1076, for example, a DVD-ROM drive, a CD-ROM drive, a DVD-RAM drive, or a CD-RAM drive can be used. In this case, the optical disk 1077 corresponding to each drive is used. A program or data can be read from the optical disk 1077 by the optical disk drive 1076 and provided to the main memory 1050 or the hard disk 1074 via the I / O controller 1070.

  The program provided to the tag determination server 10 is stored and provided in a recording medium such as the hard disk 1074, the optical disk 1077, or a memory card. The program may be installed in the tag determination server 10 and executed by being read from the recording medium via the I / O controller 1070 or downloaded via the communication I / F 1040.

  The aforementioned program may be stored in an internal or external storage medium. Here, as a storage medium constituting the storage unit 107, in addition to the hard disk 1074, the optical disk 1077, or the memory card, a magneto-optical recording medium such as an MD or a tape medium can be used. Further, a storage device such as a hard disk 1074 or an optical disk library provided in a server system connected to a dedicated communication line or the Internet may be used as a recording medium, and the program may be provided to the tag determination server 10 via the communication line.

  Here, the display device 1022 displays a screen for accepting data input to the user, or displays a screen of calculation processing results by the tag determination server 10, and is a cathode ray tube display device (CRT), a liquid crystal display device. (LCD) and other display devices.

  Here, the input means accepts input by the user, and may be configured by a keyboard, a mouse 1100, and the like.

  The communication I / F 1040 is a network adapter that enables the tag determination server 10 to be connected to a terminal via a dedicated network or a public network. The communication I / F 1040 may include a modem, a cable modem, and an Ethernet (registered trademark) adapter.

  In the above example, the tag determination server 10 has been mainly described. However, the functions described above can be realized by installing a program in a computer and operating the computer as a server device. Therefore, the functions realized by the server described as an embodiment in the present invention can be realized by executing the above-described method by the computer, or by introducing the above-mentioned program into the computer and executing it. It is.

  Further, the search server 20 and the content server 30 have the same configuration as the tag determination server 10.

[Device hardware configuration]
The terminal device 40 also has the same configuration as the tag determination server 10 described above. In the above-described example, an example realized by a so-called computer has been described. Furthermore, as long as the principle of the present invention is applicable, it can be realized by various terminals such as a mobile phone, a PDA (Personal Data Assistant), and a game machine. It's okay.

  As mentioned above, although embodiment of this invention was described, this invention is not restricted to embodiment mentioned above. The effects described in the embodiments of the present invention are only the most preferable effects resulting from the present invention, and the effects of the present invention are limited to those described in the embodiments of the present invention. is not.

It is a conceptual diagram showing the whole structure of the content management system which concerns on an example of suitable embodiment of this invention. It is a block diagram which shows each function of the tag determination server 10 which concerns on an example of suitable embodiment of this invention. It is a figure which shows the flow of the tag determination process in the tag determination server 10 which concerns on an example of suitable embodiment of this invention. It is a figure which shows the tag data table which concerns on an example of suitable embodiment of this invention. It is a figure which shows the search number table which concerns on an example of suitable embodiment of this invention. It is a figure which shows the input frequency table according to keyword which concerns on an example of suitable embodiment of this invention. It is a figure which shows the keyword ranking table which concerns on an example of suitable embodiment of this invention. It is a figure which shows the outline | summary of the similarity determination which concerns on an example of suitable embodiment of this invention. It is a figure which shows the outline | summary of the field estimation which concerns on an example of suitable embodiment of this invention. It is a figure which shows an example of the hardware constitutions of the tag determination server 10 which concerns on an example of suitable embodiment of this invention.

Explanation of symbols

DESCRIPTION OF SYMBOLS 10 Tag determination server 20 Search server 30 Content server 40 Terminal device 110 Speech recognition part 120 Morphological analysis part 130 Tag candidate extraction part 140 Search information collection part 150 Tag determination part

Claims (7)

A method in which a server adds a tag to moving image data,
Converting audio included in the moving image data into text data;
Extracting keywords from the text data;
On the extracted the keyword, in response to an inquiry to a predetermined Web search systems, statistical information indicating at least the number of times that the Web Search by the system to your Keru number of search results by the keyword or the keyword is executed acquiring search information including,
Calculating the importance of the keyword based on the acquired search information;
Adding the keyword as a tag of the moving image data when the calculated importance satisfies a predetermined condition. The method according to claim 1 , wherein the search information includes TF · IDF, which is a product of an index TF and an IDF relating to the appearance frequency of the keyword. The keyword method according to claim 1 or claim 2 which is a combination of nouns and verbs in the dependency relationship with each other. The keyword is, if it has a previously stored category data and a predetermined or more similarity, wherein the category data, to any of claims 1 to 3, step further including imparting a tag of the moving image data the method of. If the importance is less than a predetermined value, determining a field based on a predetermined rule for the keyword;
Method according to data indicating the determined the art, any steps to impart, claim 1 further comprising a according to claim 4 as a tag of said moving image data. A server for attaching a tag to moving image data,
Means for converting audio contained in the moving image data into text data;
Means for extracting a keyword from the text data;
On the extracted the keyword, in response to an inquiry to a predetermined Web search systems, statistical information indicating at least the number of times that the Web Search by the system to your Keru number of search results by the keyword or the keyword is executed Means for obtaining the search information included ;
Means for calculating the importance of the keyword based on the acquired search information;
A server provided with the keyword as a tag of the moving image data when the calculated importance satisfies a predetermined condition. A program for attaching a tag to moving image data,
Converting audio included in the moving image data into text data;
Extracting keywords from the text data;
On the extracted the keyword, in response to an inquiry to a predetermined Web search systems, statistical information indicating at least the number of times that the Web Search by the system to your Keru number of search results by the keyword or the keyword is executed acquiring search information including,
Calculating the importance of the keyword based on the acquired search information;
A program for causing a server to execute the step of assigning the keyword as a tag of the moving image data when the calculated importance satisfies a predetermined condition.

Images