JP5989157B2 - Information presenting apparatus, method, and program - Google Patents

Description

  The present invention relates to an information presentation apparatus, method, and program, and more particularly, to an information presentation apparatus, method, and program that present a search result of media data information similar to a query.

  In recent years, as the arrival of the big data era is said, the amount of data is increasing explosively. On the Internet, in addition to information transmission from companies and people familiar with computers that have been popular in the past, many other media such as SNS (social network service) that allow easy and easy information transmission have been put into practical use. Even those who do not have specialized knowledge began to send information on a daily basis. For this reason, the amount of electronic information generated on the Internet is enormous.

  Also, in the real world, as various sensors become cheaper, smaller, and more sophisticated, mobile terminals such as mobile phones and smartphones that are privately owned by many people are equipped with GPS as standard. The position information of the mobile terminal can be acquired according to the application. In addition, data such as temperature and humidity is constantly acquired by a sensor node provided for commercial use in a farm, a server room, etc., and is used for operation. As described above, sensor information acquired by various sensors is sequentially stored, accumulated, used, and the like.

  Thus, in order to utilize the enormous amount of information generated every day, it is essential to search for information, and it is important to present the search results effectively so that the user can easily understand them. For example, search and presentation of information on the Internet will be described as an example. Internet search engine is Google (registered trademark) web search, image search, YouTube (registered trademark) video search, Microsoft (registered trademark) search engine Bing (registered trademark) web search, image search, etc. As described above, the media data to be searched is wide, such as text, images, moving images, etc., but the output of search results is often made in units of several tens of items in a list format. The list format can be said to be a method of presenting search results in a one-dimensional manner in descending order of the score calculated by the search engine based on the relationship between the query and each media data. The user should examine the contents from the search results with higher ranks and cancel the confirmation of the search results when the desired information is obtained.

  However, in the presentation of search results in a one-dimensional list format, when the search key such as a keyword included in the query corresponds to not only information related to a desired target but also information related to other targets, There is a problem that a search result for a desired target and a search result for another target are mixed. For example, it is assumed that “a set of html texts on the Internet including a description relating to the person A” is searched from the web using the name of the person A as a keyword, and the list is to be obtained in descending order of score. At this time, if there is another person B who has the same name as the person A, and the “collection of html texts on the Internet including the description about the person B” has a high score, the search result list includes: "The html text on the Internet including the description regarding the person A" and "the html text on the Internet including the description regarding the person B" are mixed.

  In this case, in order to eliminate the information of another person B who has been mixed from the search results, it is necessary to add information that has A but not B as a search key to the query. It becomes necessary to analyze the characteristics of If the user does not have prior knowledge about A and B, the text shown in the list will be read and compared for characteristic analysis, and a lot of time will be consumed. In the worst case, an additional search key that is not appropriate may be selected, and information about the desired A may be excluded from the search result list.

  In order to solve the problem of presenting search results as a one-dimensional list, a method of presenting three-dimensional information has been proposed (for example, see Non-Patent Document 1).

Masatsugu Hattori et al., “Similarity Search for Picture Books Using Graph Index: Feature Fusion and Visualization of Results Graphs”, 2013 Information Processing society of Japan

  However, the information presentation method based on the three-dimensional visualization of search results described in Non-Patent Document 1 described above has not been considered as a UI for use in services.

  The present invention has been made in view of the above circumstances, and an object thereof is to provide an information presentation apparatus, method, and program capable of performing optimum information presentation that is easy for a user to use.

  In order to achieve the above object, an information presentation device according to a first aspect of the present invention provides a similarity to a query input by a user from a plurality of pieces of media data information including at least one common type of feature relating to media data. Based on the media data search unit that searches for a plurality of media data information, and the search results searched by the media data search unit, a plurality of searched media data information with high similarity to the query For each piece of media data information, a search result presentation unit that presents information including at least summary information predetermined for the media data information to the user, and a search result visualization request is input by the user. When the search result searched by the media data search unit, the plurality of searched Node indicating the respective media data information, and a graph consisting of edges that couples between media data information having similarity, is configured to include a, a visualization result presentation unit that presents to the user.

  According to the information presentation apparatus of the first aspect, the media data search unit is similar to a query input by the user from a plurality of pieces of media data information including at least one common type of feature regarding the media data. Based on the above, a plurality of media data information is searched. The search result presentation unit prioritizes media data information having a high similarity to the query as a search result searched by the media data search unit, and at least the media data information searched Information including summary information predetermined for the media data information is presented to the user. When the visualization result presentation unit receives a search result visualization request from the user, a node indicating each of the searched plurality of media data information as a search result searched by the media data search unit; and A graph composed of edges connecting between similar media data information is presented to the user.

  In this way, as a search result, priority is given to media data information having a high degree of similarity with the query, the summary information, the degree of similarity with the query, and the degree of contribution to the degree of similarity of the features having similarity with the query, When a search result visualization request is input, a graph including nodes indicating each of a plurality of searched media data information and an edge connecting between similar media data information By presenting to the user, it is possible to perform optimum information presentation that is easy for the user to use.

  The visualization result presentation unit may include the query that the plurality of searched media data information has as a search result searched by the media data search unit when a search result visualization request is input by the user. Further, each feature having similarity and the contribution of each feature to the similarity can be further presented.

The information presentation device according to the first aspect of the present invention may be configured such that, when the graph division request is input by the user, a plurality of each of the nodes is set based on similarity between media data information indicated by each of the nodes. The visualization result presentation unit is different depending on the clustering unit as a search result searched by the media data search unit when the graph division request is input by the user. Presenting the graph in which edges between the nodes separated into clusters are deleted, and further having similarity to the query included in the media data information indicated by each of the nodes belonging to the cluster for each of the plurality of clusters Each feature and the contribution of each feature to the similarity can be presented side by side.
Further, the information presentation device according to the first invention includes a clustering unit that separates each of the searched plurality of media data information into a plurality of clusters based on the similarity between the searched plurality of media data information. In addition, the search result presentation unit, as a search result searched by the media data search unit, belongs to the cluster with priority given to media data information having a high similarity to the query for each of the plurality of clusters. For each piece of media data information, information including at least summary information predetermined for the media data information can be presented to the user.

  Further, the visualization result presentation unit, when a request for dividing the graph is input by the user, as a search result searched by the media data search unit, between the nodes separated into different clusters by the clustering unit The graph in which the edge is deleted and the graph in which the display coordinates of at least one node are changed from the display coordinates of the nodes before separation by the clustering unit can be presented.

  In addition, when the graph division request is input by the user, the visualization presenting unit, as a search result searched by the media data search unit, is an edge between the nodes separated into different clusters by the clustering unit The display coordinates of at least one node are determined from the display coordinates of the nodes before separation by the clustering unit so that the display coordinates of each of the nodes are separated for each cluster. The changed graph can be presented.

  In addition, the information presentation device according to the second aspect of the present invention provides media data information based on a similarity to a query input by a user from a plurality of pieces of media data information including at least one common type of feature relating to media data. A media data search unit that searches for a plurality of media data, a node indicating each of the searched media data information, and media data information having similarity as a search result searched by the media data search unit A visualization result presentation unit that presents a graph composed of edges to the user, and when the node of the graph is designated by the user, an outline predetermined for the media data information indicated by the designated node Provide specified media data information that presents information including at least information to the user. It is configured to include a section, a.

  According to the information presentation device of the second invention, the media data search unit is similar to a query input by the user from a plurality of media data information including at least one common type of feature regarding the media data. Based on the above, a plurality of media data information is searched. The visualization result presenting unit is a search result searched by the media data search unit, a search result searched by the media data search unit, a node indicating each of the searched plurality of media data information, and similar A graph composed of edges connecting the media data information having characteristics is presented to the user. When the designated media data information presenting unit designates a node of the graph by the user, the designated media data information presenting unit provides the user with information including at least summary information predetermined for the media data information indicated by the designated node. To present.

  Thus, as a search result, a graph including a node indicating each of a plurality of searched media data information and an edge connecting between similar media data information is presented to the user. Is designated, information that includes at least the summary information of the media data information indicated by the designated node is presented to the user, so that optimum information presentation that is easy for the user to use can be performed.

  In the information presentation device according to the second invention, when a feature having similarity to the query is designated by the user, only a node indicating media data information having the designated feature can be identified. And a designated feature information presenting unit that presents the graph displaying rough information extracted from the media data information to the user.

  In addition, the designated feature information presentation unit, when a feature having similarity to the query is designated by the user, only the node indicating the media data information having the designated feature can be identified as a low information amount expression. As a rough information extracted from the media data information, the graph displaying information with a smaller information amount than the summary information can be presented to the user.

  Further, the media data can be a book.

  An information presentation method according to a third invention is an information presentation method in an information presentation device including a media data search unit, a search result presentation unit, and a visualization result presentation unit, wherein the media data search unit includes: A plurality of media data information is searched from a plurality of media data information including at least one common type of feature regarding media data based on a similarity with a query input by a user, and the search result presentation unit includes As a search result searched by the media data search unit, media data information having a high similarity to the query is given priority, and at least the media data information searched for is determined in advance for at least the media data information. Information including the displayed summary information is presented to the user, and the visualization result presentation unit provides the user When a search result visualization request is input, a search result searched by the media data search unit includes a node indicating each of the searched plurality of media data information, and between similar media data information. This is a method of presenting a graph composed of edges to be combined to the user.

  An information presentation method according to a fourth invention is an information presentation method in an information presentation device including a media data search unit, a visualization result presentation unit, and a designated media data information presentation unit, wherein the media data search unit includes: A plurality of media data information is searched from a plurality of media data information including at least one common type of feature regarding media data based on a similarity with a query input by a user, and the visualization result presentation unit includes the visualization result presentation unit, As a search result searched by the media data search unit, a graph including a node indicating each of the searched media data information and an edge connecting between the similar media data information is displayed to the user. The designated media data information presentation unit designates a node of the graph by the user. If it is, the information including at least a predetermined summary information to the media data information indicating the specified node is a method of presenting to the user.

  A program according to the fifth invention is a program for causing a computer to function as each unit constituting the information presentation device.

  As described above, according to the information presentation apparatus, method, and program of the present invention, there is an effect that it is possible to perform optimum information presentation that is easy for the user to use.

It is a block diagram which shows the structure of the information presentation apparatus which concerns on the 1st Embodiment of this invention. It is an image figure which shows an example of an input screen. It is an image figure which shows an example of the screen which presents the media data information used as a query. It is an image figure which shows an example of the media data information stored in search object media data DB. It is an image figure which shows an example of the presentation screen of a search result. It is an image figure which shows an example of the visualization result of a search result. It is the schematic for demonstrating the median centrality regarding the edge of a graph. It is an image figure which shows an example of the visualization result of a search result when isolate | separated into two clusters. It is an image figure which shows an example of the visualization result of a search result when isolate | separated into three clusters. It is an image figure which shows an example of the visualization result of the search result for demonstrating the change of the display coordinate of a node. It is an image figure which shows an example of the visualization result of a search result in the case of being isolate | separated into four clusters. It is an image figure which shows an example of the visualization result of the search result when the node of a graph is designated. It is an image figure which shows an example of the visualization result of the search result when the feature word is designated on the pie chart. It is an image figure which shows an example of the visualization result of the search result when the feature word is designated on the pie chart. It is an image figure which shows an example of the visualization result of the search result when the feature word is designated on the pie chart. It is an image figure which shows an example of the visualization result of the search result when the feature word is designated on the pie chart. It is an image figure which shows an example of the visualization result of the search result when the feature word is designated on the pie chart. It is a flowchart which shows the content of the information presentation process routine in the 1st Embodiment of this invention. It is a flowchart which shows the content of the information presentation process routine in the 1st Embodiment of this invention. It is a block diagram which shows the structure of the information presentation apparatus which concerns on the 2nd Embodiment of this invention.

  Hereinafter, embodiments of the present invention will be described in detail with reference to the drawings. In the following, each embodiment will be described with reference to a case where search target media data is “picture book”. In the following embodiments, the term “search” is used as a concept wider than “search”. Here, “search” is a task for finding media data satisfying all the conditions given as a query in the media data group to be searched. On the other hand, “search” is a task for finding a specified number of media data satisfying many conditions from all media data when there is no media data that satisfies all the conditions given as a query.

<Configuration of Information Presentation Device according to First Embodiment>
The information presentation apparatus 10 according to the first embodiment of the present invention stores a CPU (Central Processing Unit), a RAM (Random Access Memory), and an information presentation program for executing an information presentation processing routine described later. It is comprised with the computer provided with ROM (Read Only Memory). The CPU functions as the information presentation device 10 by the CPU reading and executing the information presentation program from the ROM which is an internal storage device.

  As illustrated in FIG. 1, the information presentation device 10 includes an input unit 12, a calculation unit 14, and a display unit 16.

[Input unit 12]
The input unit 12 receives an input of a query for searching for desired information on the input screen as shown in FIG. In the example of the input screen of FIG. 2, a query input is accepted by selecting media data to be a query from media data stored in a search target media data database 20 described later.

[Calculation unit 14]
The calculation unit 14 searches for a plurality of media data corresponding to the input query, and outputs a search result obtained by visualizing information indicating the searched media data on the display unit 16.

  As shown in FIG. 1, the calculation unit 14 functionally includes a search target media data database 20, a query feature extraction unit 22, a media data search unit 24, a search result presentation unit 25, a visualization result presentation unit 26, and a clustering unit. 28, a designated media data information presentation unit 30 and a designated feature information presentation unit 32.

[Query feature extraction unit 22]
The query feature extraction unit 22 receives the input query, acquires the media data information of the media data received as the query from the search target media data DB 20, as shown in FIG. As shown, the media data information of the media data to be a query is presented. The media data information to be presented may be all media data information or only a part of the title important for the user.

  Further, the query feature extraction unit 22 extracts features from the media data information of the media data received as a query, and delivers the extracted query features to the media data search unit 24.

  For example, the query feature extraction unit 22 obtains information indicating each item of the bibliographic information from the media data information of the media data given as the query, the number of occurrences of a predetermined word included in the content of the media data, and the like. It can be extracted as a feature.

The query only needs to have a feature that can be compared with the feature included in the media data information stored in the search target media data DB 20 to be described later, and exists in the data in the search target media data DB 20. May not exist. For example, if the search target media data DB 20 stores information about text included in the picture book that is the search target as a feature of the media data, the input query only needs to have text. The text may not be the expression itself that appears in the picture book stored in the search target media data DB 20.
Here, the example in which the input unit 12 receives input of media data as a query and the query feature extraction unit 22 extracts the feature from the input has been described. However, the input unit 12 has at least one feature related to media data, not media data. Input may be accepted as a query. In this case, the query feature extraction unit 22 is unnecessary, and the feature input to the input unit 12 is input to the media data search unit 24 described below. For example, the morpheme that is a feature and the number of appearances may be input to the input unit 12 as a query by the user. Further, for example, the author and title of the bibliographic information may be input. When features such as the number of appearances of morphemes and titles of bibliographic information are input, the features may be weighted from the viewpoint of normalization with those emphasized in the search or other features.
Further, the input unit 14 may accept input of data obtained by performing one paragraph processing on the media data. For example, instead of the picture book itself, an image obtained by scanning the picture book may be accepted as an input. In this case, the query feature extraction unit 22 performs OCR on the input image to extract features such as the author name.

[Media data search unit 24]
The media data search unit 24 acquires media data information having a high similarity to the query from the search target media data DB 20 using the query features passed from the query feature extraction unit 22, and the search result presentation unit 25. Then, the data is transferred to the visualization result presentation unit 26 and the clustering unit 28.

  Here, as shown in FIG. 4, the search target media data DB 20 stores a plurality of pieces of media data information. Here, the media data is an entity indicating a search target, such as a book, music, an image, and a video. The media data information is information including at least one type of feature related to the media data as described above, and can include various information such as properties and attributes indicated by the media data. For example, when the media data is an image, information indicated by each item included in metadata such as the image title, size, creation date and time, information indicating the content of the image itself, that is, image features extracted from the image data Each amount or the like can be included in the media data information indicating the image as a feature of the image.

  Here, “having the same type of feature” and “having the same feature” are different concepts. For example, it is assumed that the two media data A and B are an image having a size of 100 × 100 pixels and an image having a size of 200 × 200 pixels, respectively. Both have a plurality of “features of the same type” such as an image title, a size, and an image feature amount. However, if the “size” is noted, the value is different, so it can be said that the media data has different characteristics.

  Each media data information is given an identification number (ID) for identifying media data to be searched.

  The example of FIG. 4 is a case where the search target media data is “picture book”, and for example, appears in each item of bibliographic information such as a title of a picture book, a cover image, an author, an author of an illustration, and a publisher, and each picture book. Media data information including each of the appearance counts for each morpheme as features is stored. Here, as one of the features of media data other than bibliographic information, the number of appearances for each morpheme that appears in a picture book is used. For example, the image feature extracted from image data indicating an illustration in a picture book The volume, the voice feature amount extracted from the voice data read from the picture book, and the like may be used as the feature of the media data. Also, bibliographic information may include other bibliographic information such as a series name and picture book size as features.

  The media data search unit 24 uses the query feature extracted by the query feature extraction unit 22 and the feature included in the media data information of each media data stored in the search target media data DB 20 to use the query and each media. The similarity with the data is calculated, and the media data information of the media data with the high similarity with the query is acquired from the search target media data DB 20 in a descending order of the similarity (for example, 30).

  The calculation of the similarity between the query and the media data will be described by taking a case where the search target media data is “picture book” as an example. First, among the features included in the media data information of each picture book stored in the search target media data DB 20, the feature vector of each picture book is used by using a feature that can be compared with the query feature extracted by the query feature extraction unit 22. Create Here, each item of bibliographic information and the number of appearances for each morpheme appearing in a picture book are used as features of the media data. Specifically, the search target media data DB 20 using each item of the bibliographic information of each picture book and the number of appearances of each morpheme appearing in each picture book as characteristics of the media data stored in the search target media data DB 20. For each picture book, a vector having dimensions of the total number of types of bibliographic information and morphemes appearing in all picture books in which the media data information is stored is created. Then, a value tfidf obtained by multiplying each element of the vector by an inverse picture book appearance frequency idf (frequency of appearance tf (term frequency) in each picture book of the bibliographic information item or morpheme corresponding to the dimension of the element. Fill out. Finally, the size of the vector is normalized to 1 to obtain a feature vector for each picture book. tfidf can be calculated, for example, as shown in the following equations (1) to (3).

Here, n _{i, j} is the number of appearances of the bibliographic information item i or the morpheme i in the picture book j. D is the total number of picture books whose media data information is stored in the search target media data DB 20, d is the bibliographic information item i or morpheme i among the picture books whose media data information is stored in the search target media data DB 20 The number of picture books to be played.

  For example, media data information about a total of three picture books is stored in the search target media data DB 20 (D = 3), and the following bibliographic information items or morphemes and their morphemes are included in the media data information of each picture book. It is assumed that the feature expressed by the number of appearances is included. Note that the picture book whose ID is n is hereinafter referred to as “picture book n”. Bibliographic information is given a number of times as a weight, and the relative importance is expressed in relation to the number of morphemes. For example, if the weight of the author being X is 100 times in a book, the information that the author is X is the same weight as the appearance of a certain morpheme 100 times. The weight may be set constant for each type of the same feature (100 times for the author, 50 times for the illustration, etc.) or may be set individually.

Picture Book 1: A 3 times, B 2 times, C 2 times Picture Book 2: A 1 time, C 3 times, D 2 times Picture Book 3: A 2 times, D 1 time, E 2 times

In this case, idf _i (i = A, B, C, D, E) of each item and each morpheme is expressed by the following equation (4).

  Further, the value of each element of the vector for the picture book 1 is expressed by the following equation (5).

Finally, in order to normalize the vector size to 1, each element is divided by the square root of the sum of squares of each element. The resulting feature vector v ₁ of the picture book 1 is expressed by the following equation (6).

The feature vector v ₂ of the picture book 2 and the feature vector v 3 of the picture book ₃ are created in the same procedure.

  Next, a feature vector of the query is created. For example, it is assumed that the feature represented by the bibliographic information item or morpheme as shown below and the number of appearances thereof are extracted from the query by the query feature extraction unit 22.

      Query: A 2 times, C 3 times

In preparing a feature vector of a query, the value of idf _i in calculating tfidf utilizes values obtained from the search target media data DB 20, i.e., the expression (4). Others are obtained in the same procedure as the feature vector of the picture book as shown in the following equation (7).

Finally, in order to normalize the magnitude of the vector to 1, each element is divided by the square root of the sum of squares of each element to create a _query feature vector v _query = (0, 0, 1, 0, 0). Note that bibliographic information items or morphemes that appear in the query and that do not appear in the picture book whose media data information is stored in the search target media data DB 20 may be ignored. That is, the bibliographic information item or morpheme is not considered as an element of the query feature vector.

Next, the cosine similarity between the feature vector v _query of the created _query and each of the feature vectors v _n (n = 1, 2, 3) of each picture book is calculated. Here, since the size one feature vector is normalized to 1, cosine similarity becomes the inner product equivalent to the v _query and v _n. Also, the inner product with itself becomes 1, which is the maximum value of cosine similarity.

The media data search unit 24 acquires media data information including a feature whose similarity to the query feature is a predetermined value or more from the search target media data DB 20. Here, the feature vector v _query and cosine similarity between the feature vector v _n of each media data query in descending order, obtains the media data information pre-specified predetermined number (e.g., 30 cases). Note that the number of media data acquisitions is not limited to the upper predetermined number of similarities, and the total number of media data information stored in the search target media data DB 20 may be specified, and all media data information may be acquired in order of similarity. Good.

  In addition, the media data search unit 24 calculates the similarity between each piece of media data based on at least one of the features included in the acquired media data information. The similarity between media data can be calculated, for example, from the cosine similarity of the feature vector of each media data calculated above. The media data search unit 24 delivers the acquired media data information, the calculated similarity between media data, and the similarity between the calculated query and each media data to the search result presentation unit 25 and the visualization result presentation unit 26. .

  In addition, the similarity between each media data and the similarity between a query and each media data are not limited to the case of calculating | requiring from the similarity of each feature vector. For example, a subjective similarity or dissimilarity can be given in advance for all combinations of two media data whose media data information is stored in the search target media data DB 20. In this case, given similarity or dissimilarity can be directly used as the similarity between the media data. The similarity between the query and each piece of media data is the similarity or dissimilarity given to other media data when combined with the media data indicated by the media data information including features that match the features of the query. That's fine.

  In the present embodiment, the media data search unit 24 sets a node indicating the acquired media data information as information indicating the similarity between the media data, and the similarity between the node and the node is within a specified order for each node. Create a partial network structure that connects nodes with an edge, and prun the edges of the partial network structure that have no change in reachability when greedy search is performed even if they are deleted A network structure is created (see JP 2008-293444 A and JP 2008-305072 A). The media data search unit 24 holds the acquired media data information and the similarity between the calculated query and each media data in each node constituting the partial network structure. The media data search unit 24 delivers the partial network structure thus created to the visualization result presentation unit 26. Although the above-described partial network structure is used here, another type of partial network structure may be used as long as nodes having the same characteristics and having high similarity are connected by edges. For example, a partial network structure obtained by connecting nodes having a certain degree of similarity or higher with an edge may be used, or a partial network structure in which a node and a node whose similarity is within a specified order may be connected with an edge.

[Search result presentation unit 25]
As shown in FIG. 5, the search result presentation unit 25 searches the media data search unit 24 based on the media data information passed from the media data search unit 24 and the similarity between the media data. As a result, priority is given to the media data information having a high similarity to the query, and for each of the searched plurality of media data information (referred to as “top ranker”), at least an outline predetermined for the media data information Information including information is presented to the user by the display unit 16. For example, a screen representing the summary information, the similarity to the query, and the contribution to the similarity of each feature word having similarity to the query is presented to the user by the display unit 16. The presented information should include at least summary information predetermined for the media data information, and includes the degree of similarity to the query and the degree of contribution to the degree of similarity of each feature word having similarity to the query. There may be no other information.

  As shown in FIG. 5 above, auxiliary information related to the searched media data information is presented. The auxiliary information can include features of media data, particularly features indicating bibliographic items, similarities and ranks of queries and media data, feature words, contributions to similarity of feature words, and the like. A feature word is a word that has a high degree of contribution to being selected as media data having a high degree of similarity to a query. For example, a bibliographic item or morpheme corresponding to an element whose product between elements when a cosine similarity between a query feature vector and a media data feature vector is calculated is a feature word Can do. For the feature word, the contribution degree of the feature word is calculated based on the product between the elements corresponding to the feature word when the cosine similarity between the feature vector of the query and the feature vector of the media data is calculated.

[Visualization result presentation unit 26]
When the “graph visualization” button is pressed by the user and a search result visualization request is input by the user, the visualization result presentation unit 26 calculates a two-dimensional plane coordinate, calculates an elevation, and features as described below. Information calculation processing and visualization result generation processing are performed.

  First, the visualization result presentation unit 26 indicates each piece of media data information based on the media data information passed from the media data search unit 24 and the similarity between the pieces of media data as a two-dimensional plane coordinate calculation process. Calculate the two-dimensional plane coordinates of the node. At this time, two-dimensional such that nodes indicating media data information having high similarity between features included in the media data information are arranged close to each other, and nodes showing media data information having low similarity are arranged far from each other Calculate plane coordinates.

  In the present embodiment, the partial network structure delivered from the media data search unit 24 is used as a spring model with a fixed spring length, and the two-dimensional plane coordinates of a node indicating each piece of media data information are calculated. Specifically, in a partial network structure, nodes connected by edges are placed close to each other, nodes not connected by edges are placed far away, and all nodes included in the partial network structure are placed stably. The two-dimensional plane coordinates of each node are obtained so that

  A two-dimensional plane in which nodes indicating media data information having high similarity between features included in the media data information are arranged close to each other and nodes indicating media data information having low similarity are arranged far away from each other. Any method other than the spring model may be used as long as the method can calculate the coordinates. In addition, as described above, the two-dimensional plane coordinates are calculated using the cosine similarity between feature vectors using the number of appearances of morphemes appearing in a picture book and the features of media data represented by bibliographic items. The calculation is not limited to this, and may be performed according to the relationship between the media data information indicated by each node. For example, two-dimensional coordinates such that nodes indicating media data that matches one of the features included in the media data information (for example, “author”) are close to each other, and nodes indicating media data having different features are arranged far from each other. May be calculated.

  The visualization result presentation unit 26 passes the calculated two-dimensional plane coordinates of each node, the media data information held by each node, and the similarity between the query and each media data to the altitude calculation process.

  The visualization result presentation unit 26, as an altitude calculation process, for the node indicating each piece of media data information acquired by the media data search unit 24, the similarity between the query passed from the two-dimensional coordinate calculation process and each piece of media data. Calculate the altitude according to the degree. The altitude is height information for obtaining three-dimensional coordinates in addition to the two-dimensional plane coordinates of the nodes indicating the calculated pieces of media data information.

  As the altitude, the cosine similarity between the feature vector of each media data and the feature vector of the query may be used as it is, or the cosine similarity is multiplied by a predetermined constant (for example, 4) to obtain a height in three-dimensional coordinates. You may calculate the altitude that emphasizes. By emphasizing the height, it is possible to emphasize the contrast between the “relationship between top rankers expressed in a two-dimensional plane” and the “relationship between the top rankers expressed in the height direction and the query”. , It can make it easier to recognize top rankers that are similar (or not) to queries. In order to emphasize this contrast, the altitude may be obtained by a function in which a small value of similarity is emphasized, such as a logarithm of similarity or a square root, and a value obtained by such a function is further increased by a constant. May be multiplied.

  The visualization result presentation unit 26 displays the calculated altitude of each node, the two-dimensional plane coordinates of each node passed from the calculation process of the two-dimensional coordinates, and the similarity between the query held by each node and each media data. , And pass it to the visualization result generation process.

  The visualization result presentation unit 26 calculates, as feature information calculation processing, each feature word having a similarity with a query included in the plurality of media data information searched by the media data search unit 24 and the similarity of each feature word. The contribution is calculated. For example, a union of each feature word having similarity to a query included in a plurality of searched media data information is obtained. Further, for each feature word having similarity to the query, an average value of the contribution degree of similarity of the feature word in each of the plurality of media data information calculated by the search result presentation unit 25 is obtained, and each feature word The degree of contribution to the similarity.

  The visualization result presentation unit 26 adds the altitude calculated by the altitude calculation process to the two-dimensional plane coordinates of each node calculated by the 2D coordinate calculation process, and obtains the 3D coordinates of the node indicating each media data information. The visualization result obtained by plotting it on a three-dimensional graph is generated. The visualization result presentation unit 26 edits the media data indicated by each node so that the media data indicated by each node can be determined using at least one of the features included in the media data information indicated by each node. The right side of FIG. 6 shows an example of a visualization result when the search target media data is “picture book”. The right side of FIG. 6 shows an example of a visualization result when the three-dimensional graph is viewed from a position at a high altitude. A part of the low altitude node is displayed over the high altitude node. The right side of FIG. 6 shows an example in which a cover image of a picture book that is one of the features included in the media data information is pasted to each node. Thereby, it is possible to express which picture book each node represents. Moreover, you may use the characteristics other than a cover image, for example, you may use the text which shows the title of a picture book.

In addition, the visualization result presentation unit 26 may highlight the nodes indicating the media data having a high similarity to the query, that is, the nodes having a higher altitude by adding a darker color to the cover frame.
Moreover, although the visualization result presentation part 26 demonstrated to the case where the graph as shown in the said FIG. 6 was shown as an example, the graph to show should just be a graph which can grasp | ascertain the relationship about several picture books, It is not limited to the graph as shown in FIG.

  In addition, the visualization result presentation unit 26 contributes to each feature word having similarity to a query included in a plurality of searched media data information obtained by the feature information calculation process and the similarity of each feature word. The degree may be further presented. On the left side of FIG. 6, an example of a bar graph representing each feature word having similarity to the query and the degree of contribution to the similarity of each feature word is shown. In the bar graph, the horizontal axis represents the type of feature word, and the vertical axis represents the contribution.

[Clustering unit 28]
When a graph division request is input by the user, the clustering unit 28 separates each of the nodes into a plurality of clusters based on the similarity between the media data information indicated by each of the nodes.

  Specifically, the clustering unit 28 introduces the concept of betweenness centerality regarding the edges of the graph. The intermediary centrality regarding the edge of the graph is an idea that, for all combinations of two nodes in the graph, the edge that appears most frequently in the shortest path between the two nodes is the center in the graph. For example, taking the graph shown in FIG. 7 as an example, the shortest path between each two nodes is as follows.

AB: A → B
AC: A → C
AD: A → B → D
AE: A → B → D → E
AF: A → B → D → F
AG: A → B → D → F → G
BC: B → C
BD: B → D
BE: B → D → E
BF: B → D → F
BG: B → D → F → G
CD: C → B → D
CE: C → B → D → E
CF: C → B → D → F
CG: C → B → D → F → G
DE: D → E
DF: D → F
DG: D → F → G
EF: E → F
EG: E → G
FG: F → G

  Of these shortest paths, the most frequently appearing edge is B → D and appears 12 times. By pruning this edge, the graph shown in FIG. 7 can be separated into two clusters of A, B, and C node groups and D, E, F, and G node groups.

  However, even if one edge is removed, it may not be divided into clusters. In this case, after removing the edge, the number of appearances of the edge appearing in the shortest path in each cluster is checked again, and the edge having the maximum mediation centrality is pruned. If it still does not divide into clusters, it continues pruning the edge with the greatest median centrality until it is divided into clusters.

  When the graph division request is input by the user, the visualization result presentation unit 26 deletes the edges that have been pruned so as to be separated into clusters by the clustering unit 28 as a search result searched by the media data search unit 24. Present a graph. For example, as shown on the right side of FIG. 8, a graph separated into clusters above and below the dotted line is presented.

  Further, the visualization result presentation unit 26 has, for each cluster, each feature word having similarity to a query included in the media data information indicated by each node belonging to the cluster, and the contribution degree of each feature word to the similarity. For each cluster, each feature word having similarity to the query and the contribution degree of each feature word to the similarity are presented side by side. For example, as shown on the left side of FIG. 8, for each cluster, a bar graph representing each feature word having similarity to the query and the degree of contribution to the similarity of each feature word is presented. In the bar graph on the left side of FIG. 8, it can be seen that the cluster having the feature word “XXΔΔko” as the author information is separated from the cluster having no author word. In this way, when a graph is divided in a network structure in which nodes with high similarity due to having the same features are connected by edges, clusters formed by those having the same features are obtained.

  The clustering unit 28 further separates each of the nodes into clusters based on the similarity between the media data information indicated by each of the nodes each time a graph division request is further input by the user.

  The visualization result presentation unit 26 is pruned so as to be separated into clusters by the clustering unit 28 as a search result searched by the media data search unit 24 every time a graph division request is further input by the user. Present a graph with edges removed. For example, as shown on the right side of FIG. 9, a graph separated into three clusters is presented. Also, as shown on the left side of FIG. 9, for each of the three clusters, a bar graph representing each feature word having similarity to the query and the degree of contribution to the similarity of each feature word is presented.

  In addition, when the graph division request is input by the user, the visualization result presentation unit 26 uses the search result searched by the media data search unit 24 to display the edges pruned so as to be separated into clusters by the clustering unit 28. A graph that is a deleted graph and in which the display coordinates of at least one node are changed from the display coordinates of the node before separation by the clustering unit 28 so that the display coordinates of each node are separated for each cluster is presented. To do. For example, the center coordinates of the cluster are adjusted by changing the display coordinates of the nodes constituting the cluster so that the display coordinates of each node are separated for each cluster. As a result, as shown on the right side of FIG. 10, a graph easy to understand that it is divided into three clusters is presented. In the example of FIG. 10, the length of the edge is adjusted so that the nodes are not covered and displayed. In the example of FIG. 10, it can be seen that the upper large cluster is divided into two clusters indicating whether or not the feature word “□□ bookstore” is included as publisher information.

  Further, when a graph division request is further input by the user, a graph separated into four clusters is presented as shown on the right side of FIG. Also, as shown on the left side of FIG. 11, for each of the four clusters, a bar graph representing each feature word having similarity to the query and the degree of contribution to the similarity of each feature word is presented.

  In FIG. 11, each of the three clusters has a feature word “□□ bookstore” as publisher information, a feature word “wolf”, and a feature word “XXXXX” as author information. It can be seen that it accounts for the majority of the degree. In another cluster, the contributions of the feature words “AAA”, “Grandma”, and “Wolf” are relatively high.

  While performing the graph division step by step, the user determines that the graph division has been performed appropriately, and ends the graph division. Note that, after the graph division is completed, or in the middle of the graph division, as shown in FIG. 5, a screen displaying a list of searched media data information may be displayed again. Further, in the present embodiment, the bar graph representing the features for each cluster on the left side of FIG. 11 has been described using an example limited to the feature words that the query and the top ranker have in common. Words may be used together. When the specified number of search top rankers is large or the media data similar to the query is small in the media data DB, the lower top rankers belong if only the feature words that the query and the top rankers have in common are used. A cluster contributes less to any feature word, making it difficult for the user to recognize the commonality of the cluster. If a feature word that is not possessed by a query and is possessed only by the top ranker is used in combination, the meaning of the cluster is facilitated by the feature word.

[Designated media data information presentation unit 30]
When the node on the graph is designated by the user and the media data information is designated by the user, the designated media data information presenting unit 30 is configured to obtain summary information predetermined for the media data information indicated by the designated node, a query, The display unit 16 presents the user with at least one of the similarity and the contribution of the feature word having similarity to the query included in the media data information to the similarity.

  For example, as shown on the left side of FIG. 12, auxiliary information related to designated media data information is presented. The auxiliary information may include features of the media data, in particular, each item of the bibliographic information, similarity and ranking between the query and each media data, feature words, a pie chart showing the contribution of each feature word to the similarity, etc. it can.

[Designated feature information presentation unit 32]
The designated feature information presentation unit 32 is a media having the designated feature word when the feature word is designated by the user on the pie chart representing the contribution degree of each feature word presented by the designated media data information presentation unit 30. Only the node indicating the data information is information that can be identified as a low information amount expression, and a graph displaying rough information extracted from the media data information is presented to the user by the display unit 16. Information that is an identifiable low information amount expression and that has a smaller information amount than the summary information of the media data information may be displayed as rough information extracted from the media data information. For example, FIGS. As shown in FIG. 17, only the node indicating the picture book having the specified feature word is presented with a graph displaying the cover of the picture book. In the example of FIGS. 13 and 14, the feature word “AAA” is designated, and the cover of the picture book is displayed only at the node indicating the picture book having the feature word “AAA”. In the example of FIG. 15, the feature word “wolf” is designated, and the cover of the picture book is displayed only for the node indicating the picture book having the feature word “wolf”. In the example of FIG. 16, the feature word “□□ bookstore” as the publisher information is specified, and the cover of the picture book is displayed only in the node indicating the picture book having the feature word “□□ bookstore” as the publisher information. In the example of FIG. 17, the feature word “XXΔΔko” as the author information is designated, and the cover of the picture book is displayed only for the node indicating the picture book having the feature word “XXΔΔko” as the author information. Yes.

<Operation of Information Presentation Device 10 according to First Embodiment>
Next, the operation of the information presentation device 10 according to the present embodiment will be described. First, on the input screen shown in FIG. 2, when the media data information to be a query is selected by the user and the query input is accepted, the information presentation apparatus 10 executes the information presentation processing routine shown in FIGS. The

  First, in step S100, the query feature extraction unit 22 receives the input query, acquires the media data information selected as the query from the search target media data DB 20, and displays the display unit 16 as shown in FIG. Presents the media data information selected as a query. When the execution of the similarity search is instructed, the process proceeds to step S102.

  In step S <b> 102, the query feature extraction unit 22 extracts features from the media data information serving as a query, and passes the extracted query features to the media data search unit 24.

  Next, in step S104, the media data search unit 24 acquires media data information having a high similarity to the query from the search target media data DB 20 using the query features passed from the query feature extraction unit 22. To do. For example, the media data search unit 24 includes a feature vector created from the query features extracted by the query feature extraction unit 22 and features included in the media data information of each media data stored in the search target media data DB 20. The cosine similarity between the query feature and the feature vector created from the comparable feature is calculated as the similarity between the query and each media data, and the similarity between the query and the feature vector is calculated from the search target media data DB 20. Media data information having a high level is acquired in a predetermined number (for example, 30) in descending order of similarity.

  Then, the media data search unit 24 calculates the similarity between the media data based on at least one of the features included in the acquired media data information, and obtains between the acquired media data information and the calculated media data. The similarity and the similarity between the calculated query and each media data are transferred to the search result presentation unit 25 and the visualization result presentation unit 26. For example, the media data search unit 24 sets a node indicating the acquired media data information based on the similarity between the features of the acquired media data, and the similarity between the features included in the media data information indicated by each node Are connected to each other by an edge, and a partial network structure in which the similarity between the media data information and the query and each media data is maintained in each node is created and transferred to the visualization result presentation unit 26 .

  Next, in step S106, the search result presentation unit 25 determines that the media data based on the media data information passed from the media data search unit 24 and the similarity between the media data as shown in FIG. As a search result searched for by the search unit 24, summary information predetermined for the media data information for each of a plurality of searched media data information by giving priority to media data information having a high similarity to the query. The display unit 16 presents a screen representing the similarity to the query and the contribution of the feature word having similarity to the query to the similarity.

  In step S108, it is determined whether a search result visualization request is input by the user. When the “graph visualization” button is pressed by the user and a search result visualization request is input, the process proceeds to step S110.

  In step S110, the visualization result presentation unit 26 calculates each piece of media data information based on the media data information passed from the media data search unit 24 and the similarity between the pieces of media data as a calculation process of the two-dimensional plane coordinates. The two-dimensional plane coordinates of the indicated node are arranged close to each other indicating the media data information having a high similarity between the features included in the media data information, and distant from each other indicating the media data information having a low similarity. Calculate as follows. For example, the visualization result presentation unit 26 uses the partial network structure delivered from the media data search unit 24 as a spring model with a fixed spring length, and nodes connected by edges are arranged close to each other and are not connected by edges. The nodes are arranged far from each other, and the two-dimensional plane coordinates of each node are obtained so that all the nodes included in the partial network structure are stably arranged. The visualization result presentation unit 26 passes the calculated two-dimensional plane coordinates of each node, the media data information held by each node, and the similarity between the query and each media data to the altitude calculation process.

  Next, in step S112, the visualization result presentation unit 26, for the node indicating each piece of media data information acquired by the media data search unit 24, the query passed from the calculation process of the two-dimensional plane coordinates, each piece of media data, The altitude corresponding to the similarity is calculated. The visualization result presentation unit 26 receives the calculated elevation of each node, the two-dimensional plane coordinates of each node, the media data information held by each node, and the similarity between the query and each media data, to the visualization result generation process. hand over.

  Then, in step S113, the visualization result presenting unit 26, as the feature information calculation process, includes each feature word having similarity to the query included in the plurality of media data information searched by the media data search unit 24, and the similarity degree. The contribution of each feature word to is calculated.

  Next, in step S114, the visualization result presentation unit 26 adds the elevation calculated in step S112 to the two-dimensional plane coordinates of each node calculated in step S110 as a visualization result generation process, A three-dimensional coordinate of a node indicating data information is obtained, and a visualization result is generated by plotting this on a three-dimensional graph. The visualization result presentation unit 26 edits the media data indicated by each node so that the media data indicated by each node can be determined using at least one of the features included in the media data information indicated by each node. For example, when the search target media data is “picture book”, the cover image of the picture book is pasted to each node. In addition, the visualization result presentation unit 26 includes each feature word having similarity to the query included in the searched plurality of media data information obtained in step S113, and the degree of contribution to the similarity of each feature word. Is added so that the visualization result is presented by the display unit 16.

  In step S116, it is determined whether or not a graph division request is input by the user. When a graph division request is input by the user, the process proceeds to step S118.

  In step S118, the clustering unit 28 introduces the concept of betweenness centerality regarding the edges of the graph, and prunes edges based on the similarity between the media data information indicated by each node. Each node is separated into a plurality of clusters. Then, the visualization result presentation unit 26 generates a visualization result reflecting the result of separation into a plurality of clusters by the clustering unit 28. In addition, the visualization result presentation unit 26 generates a bar graph that reflects the result of being separated into a plurality of clusters by the clustering unit 28. The visualization result presentation unit 26 adds a bar graph and outputs the visualization result so that the display unit 16 presents the visualization result.

  In step S120, it is determined whether a graph division request is input by the user. When the graph division request is input by the user, the process returns to step S118. If the user does not input a graph division request, the process proceeds to step S122.

  In step S122, it is determined whether or not a node on the graph is designated by the user. When a node on the graph is designated by the mouse cursor according to the user operation, the process proceeds to step S124. On the other hand, if the node on the graph is not designated by the user, the process returns to step S120.

  In step S124, the designated media data information presenting unit 30 uses the similarity between the outline information predetermined for the media data information indicated by the designated node, the query, and the query similarity that the media data information has. The display unit 16 presents at least one or more of each feature word having a pie chart representing the degree of contribution to the similarity to the user.

  Next, in step S126, it is determined whether or not a feature word is designated on the pie chart by the user. When a feature word is designated on the pie chart with the mouse cursor according to the user's operation, the process proceeds to step S128.

  In step S128, the designated feature information presenting unit 32 is information that can be identified as a low information amount expression only for the node indicating the media data information having the designated feature word, and is coarsely extracted from the media data information. The graph displaying the information is presented to the user by the display unit 16.

  In step S130, it is determined whether or not to end the presentation of the search result. If not, the process returns to step S126. On the other hand, when the end of the presentation of the search result is instructed by the user operation, the information presentation processing routine is ended.

  As described above, according to the information presenting apparatus according to the present embodiment, as a search result, priority is given to the media data information having a high similarity to the query, the summary information, the similarity to the query, and the similarity. When the search result visualization request is input to the user, the degree of contribution of the feature word to the degree is presented to the user, and the media data having similarities and the media data having similarity By presenting a graph composed of edges connecting information to the user, it is possible to present optimal information that is easy for the user to use.

  Also, as a search result, a graph including nodes indicating each of a plurality of searched media data information and an edge connecting between similar media data information is presented to the user, and the node of the graph is designated. If it is, the outline information of the media data information indicated by the designated node, the similarity to the query, and the contribution to the similarity of the feature words of the media data information are presented to the user. Thus, it is possible to provide optimum information that is easy for the user to use.

  Also, unlike list-type result display, graph-type result display makes it easy to understand the relationship between media data with high similarity to queries, and as a result, unlike media data search using collaborative filtering. Information can be presented so that the media data is recommended (same author, same morpheme).

  In addition, it is easy to eliminate noise that is not necessary for the user, which is difficult if the contents are not read in the list type result display.

  The present invention is not limited to the above embodiment, and various modifications and applications can be made without departing from the gist of the present invention.

  For example, although the case where the media data to be searched is a picture book has been described as an example, the present invention is not limited to this. Media data that does not change in the time direction, such as text and images, like a picture book, or media data that is time-series data including video may be used. It may also be audio data.

  Further, in the case where the media data is a picture book, the case where the cover is displayed as information having a smaller amount of information than the summary information of the media data shown in the graph node has been described as an example. However, the present invention is not limited to this. For example, when the media data is a book (book), a cover or a combination of a cover and a title character string may be displayed on a node of the graph. When the media data is a CD, a jacket cover may be displayed at a node of the graph. When the media data is a video, a thumbnail image may be displayed at a node of the graph, or a display with movement such as gif animation may be used. When the media data is speech, a speaker icon (for example, a male or female icon), a group icon, or a musical instrument icon may be displayed at a node of the graph. Further, when the media data is a newspaper article, a title or an article number may be displayed at a node of the graph.

  In addition, an example has been described in which the number of clusters is increased by one each time a graph division request is input. However, the present invention is not limited to this, and it is automatically performed until a predetermined condition is satisfied. Cluster separation may be repeated. For example, a technique called modularity maximization (reference non-patent document 1 “Newman, M. E. J.“ Modularity and community structure in networks ”. PROCEEDINGS-NATIONAL ACADEMY OF SCIENCES USA 103 (23), 2006”) can be used. In this method, the modularity Q is a function that is intuitively “a larger value when there are more edges from one node to a node belonging to the same cluster than to a node belonging to another cluster”. May be introduced each time an edge is pruned, and Q may be examined, and an edge having the maximum mediation centrality may be pruned as many times as Q takes the maximum value.

<Configuration of Information Presentation Device according to Second Embodiment>
Next, an information presentation apparatus according to the second embodiment will be described. In addition, about the part which becomes the structure similar to 1st Embodiment, the same code | symbol is attached | subjected and description is abbreviate | omitted.

[Calculation unit 214]
As illustrated in FIG. 20, the calculation unit 214 of the information presentation device 200 according to the second embodiment functionally includes a search target media data database 20, a query feature extraction unit 22, a media data search unit 24, clustering. It can be expressed by a configuration including a unit 228, a search result presentation unit 225, a visualization result presentation unit 226, a designated media data information presentation unit 30, and a designated feature information presentation unit 32.

[Media data search unit 24]
The media data search unit 24 acquires media data information having a high similarity to the query from the search target media data DB 20 using the query features passed from the query feature extraction unit 22, and the clustering unit 228, search The result is transferred to the result presentation unit 225 and the visualization result presentation unit 226.

[Clustering unit 228]
The clustering unit 228 searches the media data based on the similarity between the searched media data information based on the media data information passed from the media data search unit 24 and the similarity between the media data. Separate each piece of information into multiple clusters. The clustering unit 228 automatically repeats the separation of the clusters until the predetermined condition is satisfied by automatically repeating the separation of the clusters until the predetermined condition is satisfied. As mentioned above, for example, modularity Q is introduced by the method called modularity maximization (Reference Non-Patent Document 1 “Newman, MEJ“ Modularity and community structure in networks ”. PROCEEDINGS- NATIONAL ACADEMY OF SCIENCES USA 103 (23), 2006”). Then, Q is checked every time an edge is pruned, and a method of pruning an edge having the maximum median centrality as many times as Q takes a maximum value is used.

[Search result presentation unit 225]
The search result presentation unit 225 searches the media data search unit 24 based on the media data information passed from the media data search unit 24, the clustering result by the clustering unit 228, and the similarity between the media data. As a result, for each piece of media data information belonging to the cluster, information including at least summary information predetermined for the media data information is given priority over media data information having a high similarity to the query. Presented to the user by the display unit 16. For example, a screen representing the summary information, the similarity to the query, and the contribution to the similarity of each feature word having similarity to the query is presented to the user by the display unit 16.

[Visualization result presentation unit 226]
The visualization result presentation unit 226 receives a search result visualization request from the media data search unit 24 when a “graph visualization” button provided for each cluster is pressed by the user and a search result visualization request is input. Based on the media data information having a high degree of similarity with the query, a graph capable of grasping the relationship between a plurality of picture books is generated and presented in the same manner as in the first embodiment. For example, as described in the first embodiment, the media data information passed from the media data search unit 24 and the similarity between each media data are input, An altitude calculation process, a feature information calculation process, and a visualization result generation process are performed to generate and present a graph of media data information belonging to the corresponding cluster.

  Also, the visualization result presentation unit 226, when the “graph visualization” button corresponding to all the clusters is pressed by the user and a search result visualization request is input, in the same way as in the first embodiment, Based on the media data information having a high degree of similarity with the query passed from the data search unit 24, a graph that can grasp the relationship between a plurality of picture books is generated as in the first embodiment. To present. For example, as described in the first embodiment above, the media data information passed from the media data search unit 24 and the similarity between each media data are input, and a calculation process of two-dimensional plane coordinates, An altitude calculation process, a feature information calculation process, and a visualization result generation process are performed, and a graph from which edges that have been pruned so as to be separated into clusters by the clustering unit 228 are generated is generated and presented. For example, the graph shown on the right side of FIG. 11 is presented.

  In addition, since it is the same as that of 1st Embodiment about the other structure and effect | action of the information presentation apparatus 200 which concern on 2nd Embodiment, description is abbreviate | omitted.

The present invention is not limited to the above-described embodiment, and various modifications and applications are possible without departing from the gist of the present invention.

  For example, in the same way as in the first embodiment, when a graph division request is further input by the user, each of the nodes is further clustered based on the similarity between the media data information indicated by each of the nodes. You may make it present the graph which remove | separated and deleted the edge pruned so that it might isolate | separate into a cluster. In this case, as in the first embodiment, the display coordinates of at least one node are the display coordinates of the node before separation by clustering so that the display coordinates of each node are separated for each cluster. You may make it show the graph which changed from.

  Further, in the present specification, the embodiment in which the program is preinstalled has been described. However, the program stored in an external storage device or recording medium is read as needed, and is downloaded and executed via the Internet. You may make it do. In addition, the program can be provided by being stored in a computer-readable recording medium.

10, 200 Information presentation device 12 Input unit 14, 214 Calculation unit 16 Display unit 20 Search target media data database 22 Query feature extraction unit 24 Media data search unit 25, 225 Search result presentation unit 26, 226 Visualization result presentation unit 28, 228 Clustering unit 30 Designated media data information presentation unit 32 Designated feature information presentation unit

Claims (14)

A media data search unit for searching a plurality of media data information based on a similarity to a query input by a user from a plurality of media data information including at least one common type of feature regarding the media data;
As a search result searched by the media data search unit, media data information having a high similarity to the query is given priority, and for each of the searched media data information, at least the media data information is previously stored. A search result presentation unit for presenting information including the defined summary information to the user;
When a search result visualization request is input by the user, as a search result searched by the media data search unit, a node indicating each of the searched plurality of media data information, and media data information having similarity A graph composed of edges that join each other, each feature having similarity to the query included in the searched plurality of media data information, and a contribution degree of each feature to the similarity to the user. A visualization result presentation unit to present
An information presentation device. A media data search unit for searching a plurality of media data information based on a similarity to a query input by a user from a plurality of media data information including at least one common type of feature regarding the media data;
As a search result searched by the media data search unit, media data information having a high similarity to the query is given priority, and for each of the searched media data information, at least the media data information is previously stored. A search result presentation unit for presenting information including the defined summary information to the user;
When a search result visualization request is input by the user, as a search result searched by the media data search unit, a node indicating each of the searched plurality of media data information, and media data information having similarity A visualization result presenting unit that presents a graph composed of edges that join between them to the user;
A clustering unit that separates each of the nodes into a plurality of clusters based on the similarity between the media data information indicated by each of the nodes when the graph division request is input by the user;
When the graph division request is input by the user, the visualization result presentation unit displays an edge between the nodes separated into different clusters by the clustering unit as a search result searched by the media data search unit. The deleted graph is presented, and for each of the plurality of clusters, the media data information indicated by each of the nodes belonging to the cluster has features similar to the query, and features to the similarity An information presentation device that presents the contributions side by side . A clustering unit that separates each of the searched plurality of media data information into a plurality of clusters based on the similarity between the searched plurality of media data information;
The search result presentation unit gives priority to media data information having a high similarity to the query for each of the plurality of clusters as the search result searched by the media data search unit, and media data information belonging to the cluster for each, information, information presentation apparatus according to claim 1, wherein presenting to the user with a summary information that is predetermined for at least the media data information. When the graph division request is input by the user, the visualization result presentation unit displays an edge between the nodes separated into different clusters by the clustering unit as a search result searched by the media data search unit. a deleted the graph, and, at least the display coordinates of the one node, the information presentation apparatus according to claim 2 or 3, wherein presenting the graph changes from the display coordinates of the node before separation by the clustering unit. A media data search unit for searching a plurality of media data information based on a similarity to a query input by a user from a plurality of media data information including at least one common type of feature regarding the media data;
A clustering unit for separating each of the searched plurality of media data information into a plurality of clusters based on the similarity between the searched plurality of media data information;
As a search result searched by the media data search unit, for each of the plurality of clusters, the media data information having a high similarity with the query is prioritized and the media data information belonging to the cluster is searched. For each, a search result presentation unit that presents information including at least summary information predetermined for the media data information to the user;
When a search result visualization request is input by the user, as a search result searched by the media data search unit, a node indicating each of the searched plurality of media data information, and media data information having similarity A visualization result presenting unit that presents a graph composed of edges that join between them to the user;
Only including,
When the graph division request is input by the user, the visualization result presentation unit displays an edge between the nodes separated into different clusters by the clustering unit as a search result searched by the media data search unit. The deleted graph and the graph in which the display coordinates of at least one node are changed from the display coordinates of the nodes before separation by the clustering unit are presented.
Information presentation device. When the graph division request is input by the user, the visualization result presentation unit displays an edge between the nodes separated into different clusters by the clustering unit as a search result searched by the media data search unit. In the deleted graph, the display coordinates of at least one node change from the display coordinates of the node before separation by the clustering unit so that the display coordinates of each of the nodes are separated for each cluster. The information presentation apparatus according to claim 4 or 5 , wherein the graph is presented. A media data search unit for searching a plurality of media data information based on a similarity to a query input by a user from a plurality of media data information including at least one common type of feature regarding the media data;
As a search result searched by the media data search unit, a graph including a node indicating each of the searched media data information and an edge connecting between the similar media data information is displayed to the user. A visualization result presentation unit to present
When the node of the graph is designated by the user, the designated media data information presentation that presents to the user information including at least summary information predetermined for the media data information indicated by the designated node And
When a feature having similarity to the query is designated by the user, only the node indicating the media data information having the designated feature is information that can be identified as a low information amount expression, and the media A designated feature information presentation unit for presenting the graph displaying rough information extracted from data information to the user;
An information presentation device. When a feature having similarity to the query is designated by the user, the designated feature information presenting unit provides the identifiable low information amount expression only for the node indicating the media data information having the designated feature. The information presenting apparatus according to claim 7 , wherein the graph displaying information that is information and has less information than the summary information is presented to the user as rough information extracted from the media data information. The media data, the present and claims 1 information presentation apparatus according to any one of claims 8. An information presentation method in an information presentation device including a media data search unit, a search result presentation unit, and a visualization result presentation unit,
The media data search unit searches a plurality of media data information based on a similarity with a query input by a user from a plurality of media data information including at least one common type of feature regarding the media data,
The search result presentation unit prioritizes media data information having a high similarity to the query as a search result searched by the media data search unit, and at least the media data information searched Presenting the user with information including summary information predetermined for the media data information,
When the visualization result presentation unit receives a search result visualization request from the user, a node indicating each of the searched plurality of media data information as a search result searched by the media data search unit; and A graph composed of edges connecting between media data information having similarity, each feature having similarity with the query included in the searched plurality of media data information, and contribution of each feature to the similarity steal, information presentation method for presenting to the user. An information presentation method in an information presentation device including a media data search unit, a search result presentation unit, a visualization result presentation unit, and a clustering unit ,
The media data search unit searches a plurality of media data information based on a similarity with a query input by a user from a plurality of media data information including at least one common type of feature regarding the media data,
The search result presentation unit prioritizes media data information having a high similarity to the query as a search result searched by the media data search unit, and at least the media data information searched Presenting the user with information including summary information predetermined for the media data information,
When the visualization result presentation unit receives a search result visualization request from the user, a node indicating each of the searched plurality of media data information as a search result searched by the media data search unit; and Presenting to the user a graph consisting of edges connecting between similar media data information ;
When the clustering unit receives a request to divide the graph by the user, based on the similarity between the media data information indicated by each of the nodes, each of the nodes is divided into a plurality of clusters,
When the visualization result presentation unit receives a request to divide the graph by the user, as a search result searched by the media data search unit, an edge between the nodes separated into different clusters by the clustering unit is displayed. The deleted graph is presented, and for each of the plurality of clusters, the media data information indicated by each of the nodes belonging to the cluster has features similar to the query, and features to the similarity Information presentation method that presents the contributions of each side by side . An information presentation method in an information presentation device including a media data search unit, a search result presentation unit, a visualization result presentation unit, and a clustering unit ,
The media data search unit searches a plurality of media data information based on a similarity with a query input by a user from a plurality of media data information including at least one common type of feature regarding the media data,
The clustering unit separates each of the searched media data information into a plurality of clusters based on the similarity between the searched media data information,
The search result presenting unit is searched as a search result searched for by the media data search unit with priority given to media data information having a high similarity to the query for each of the plurality of clusters , and For each piece of media data information belonging to the cluster , at least information including summary information predetermined for the media data information is presented to the user,
When the visualization result presentation unit receives a search result visualization request from the user, a node indicating each of the searched plurality of media data information as a search result searched by the media data search unit; and Presenting to the user a graph consisting of edges connecting between similar media data information ;
When the visualization result presentation unit receives a request to divide the graph by the user, as a search result searched by the media data search unit, an edge between the nodes separated into different clusters by the clustering unit is displayed. The deleted graph and the graph in which the display coordinates of at least one node are changed from the display coordinates of the nodes before separation by the clustering unit are presented.
Information presentation method. An information presentation method in an information presentation device including a media data search unit, a visualization result presentation unit, a designated media data information presentation unit, and a designated feature information presentation unit ,
The media data search unit searches a plurality of media data information based on a similarity with a query input by a user from a plurality of media data information including at least one common type of feature regarding the media data,
The visualization result presentation unit includes, as a search result searched by the media data search unit, a node indicating each of the plurality of searched media data information and an edge connecting between the similar media data information Presenting a graph to the user;
When the designated media data information presenting unit designates a node of the graph by the user, the designated media data information presenting unit provides the user with information including at least summary information predetermined for the media data information indicated by the designated node. It presents for,
When the designated feature information presenting unit designates a feature having similarity to the query by the user, only the node indicating the media data information having the designated feature can be identified as a low information amount expression. And the information presentation method which presents the said graph which displayed the rough information extracted from the said media data information with respect to the said user . A program for causing a computer to function as each section of the information presentation apparatus according to any one of claims 1 to 9.