JP2006228162A - Information processor and program - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To detect changes with time of data having a graph structure constituted of nodes and links by solving the problem that it is impossible to segment a partial graph concerning relations between nodes, which change with time, in conventional graph structure data processing techniques. <P>SOLUTION: This information processor includes; a link strength information acquisition part for acquiring node information and link strength information being information showing frequencies of interactions between nodes per prescribed time; a partial graph determination part for determining whether a partial graph constituted of two or more nodes and links between these nodes meets a condition relating to a prescribed degree of coupling or not on the basis of acquired one or more pieces of link strength information; a partial graph information acquisition part for acquiring partial graph information being information relating to the partial graph, in the case that the partial graph is determined to meet the prescribed condition; and a partial graph information processing part for performing processing on the basis of the partial graph information. Changes with time of data having the graph structure constituted of nodes and links can be detected by the information processor. <P>COPYRIGHT: (C)2006,JPO&NCIPI

Description

  The present invention relates to an information processing apparatus that extracts a partial graph, which is a partial graph, from a temporal change in the degree of connection of links that connect nodes constituting graph data having a graph structure.

  In a conventional moving image summarizing method for summarizing moving images and the like, there is a moving image summarizing method capable of automatically extracting semantically important events in a moving image (for example, see Patent Document 1). The video summarization method includes a feature extractor 40 that extracts video features, a probability model 42 that uses a model such as a hidden Markov model to integrate features and determine boundaries, and commercial and non-commercial slow motion playback segments. A commercial / non-commercial filter 44 and a summary generator 46 that generates a summary based on the detected slow motion playback segment. The feature extractor 40 in this moving image summarization method extracts features from the color histogram in block 50 and extracts three features from the pixel-based differences in block 52. The features extracted at block 52 characterize each component of the playback segment's slow motion, still field, and / or normal speed playback. The feature extracted in block 50 in the moving image summarizing method characterizes the editing effect component.

In addition, research to develop an algorithm for extracting a set of highly related nodes from a graph structure has been conventionally performed (see Non-Patent Document 2 and Non-Patent Document 3). Such research detects, for example, universal communities that do not change with time, such as Web communities and social networks, and compresses or eliminates time components when handling time-dependent information. To estimate the community.
Further, Non-Patent Document 1 discloses a method for acquiring content to be given to an information processing apparatus that performs video summarization described in this embodiment.
JP 2002-232840 A (first page, FIG. 1 etc.) Yasuyuki Kaku et al., “Recording and Sharing Experiences in Ubiquitous Environment”, System / Control / Information (Journal of System Control Information Society), November 2004, Vol. 48, no. 1, pp. 458-463 J. et al. Kleinberg. "Authoritative sources in a hyperlinked environment", Proc. 9th ACM-SIAM Symposium on Discrete Algorithms, 1998 Chris Ding, et al., “Special Min-Max Cut for Graph Partitioning and Data Clustering.”, Proc. of 1st IEEE Int Conf. Data Mining. San Jose, CA, 2001. pp. 107-114

However, in the information processing apparatus that implements the conventional moving image summarizing method, since attention is not paid to the interaction between people and objects appearing in the moving image (hereinafter, collectively referred to as “object”), an appropriate moving image is used. There was a problem that could not be summarized.
In addition, according to the data processing research of the conventional graph structure, the concept of time has not been introduced, and it has been impossible to cut out a partial graph focusing on the relationship between nodes that changes with the passage of time.

  The information processing apparatus according to the first aspect of the present invention is a link strength information acquisition unit that acquires node information that is information relating to a node, and link strength information that is information indicating the frequency of interaction per predetermined time between the nodes, A portion that determines whether or not a subgraph composed of two or more nodes and links between the nodes satisfies a condition related to a predetermined degree of connection based on one or more link strength information acquired by the link strength information acquisition unit A graph determination unit; and a partial graph information acquisition unit that acquires partial graph information that is information about the partial graph when the partial graph determination unit determines that the partial graph satisfies the predetermined condition; The information processing apparatus includes a partial graph information processing unit that performs processing based on graph information.

  In addition, the information processing apparatus according to the second aspect of the invention calculates the degree of coupling of the subgraph based on the one or more link strength information acquired by the link strength information acquisition unit with respect to the first invention, A partial graph connection strength information acquisition unit configured to acquire graph connection strength information, wherein the partial graph determination unit determines whether the partial graph has a predetermined degree of connection based on the link strength information and the partial graph connection strength information; It is the information processing apparatus which judges whether the conditions regarding are satisfy | filled.

  With such a configuration, it is possible to detect temporal changes in data having a graph structure composed of nodes and links, detect temporally changing subgraphs (graphs formed by combining nodes), and make meaningful meaningful subgraphs. Information can be acquired.

Further, the information processing apparatus according to the third aspect of the present invention is based on the link strength information between one node and other nodes of the two or more nodes constituting the subgraph. The subgraph determination unit further includes a connection contribution information acquisition unit that calculates a connection contribution that is information indicating a connection contribution in the subgraph of the one node and acquires connection contribution information. , Based on the link strength information and the coupling contribution information, or the link strength information and the subgraph coupling strength information and the coupling contribution information, whether or not the partial graph satisfies a condition regarding a predetermined coupling degree. It is an information processing device to judge.
With this configuration, it is possible to detect a state in which a node is to be deleted from a subgraph and to construct a new subgraph.

Further, in the information processing apparatus according to the fourth aspect of the present invention, in contrast to the first aspect, the link strength information acquisition unit is node information that does not exist in node information included in one or more stored partial graph information. Link strength information between nodes corresponding to the information, and the partial graph determination unit determines whether the link strength information acquired by the link strength information acquisition unit is information indicating a degree of coupling equal to or higher than a predetermined value. It is a processing device.
With this configuration, generation of a new subgraph can be detected.

Further, in the information processing apparatus according to the fifth aspect of the invention, in contrast to the second aspect of the invention, the link strength information acquisition unit is one node information in node information included in one or more stored partial graph information. Link strength information between the first node corresponding to the second node corresponding to the node information not existing in the node information included in the one or more stored partial graph information, and the portion When the link strength information acquired by the link strength information acquisition unit and the partial graph connection strength information of the subgraph having the first node satisfy a predetermined condition, the graph determination unit adds the second graph to the second subgraph. This is an information processing apparatus that determines that the subgraph to which the node is added satisfies a condition regarding a predetermined degree of connectivity.
With this configuration, the addition of a new node to the subgraph can be detected.

Further, in the information processing apparatus according to the sixth aspect of the invention, in contrast to the second aspect of the invention, the link strength information acquisition unit stores the node information in the node information included in the stored partial graph information. The link strength information between the corresponding first node and the second node corresponding to one node information in the node information included in the other stored partial graph information is acquired, and the partial graph determination is performed. The link strength information acquired by the link strength information acquisition unit and the partial graph connection strength information of the first subgraph corresponding to the one partial graph information satisfy a predetermined condition, and the link strength information acquisition unit acquires Determined whether the subgraph connection strength information of the second subgraph corresponding to the link strength information and the other subgraph information satisfies a predetermined condition, and the subgraph determining unit determines the predetermined condition. If it is determined that the link strength information is acquired, the link strength information acquisition unit is instructed to acquire link strength information between nodes constituting the first subgraph and the second subgraph, or the link strength information acquisition unit Instructing to acquire link strength information between nodes constituting the first subgraph and the second subgraph and instructing the subgraph connection strength information acquisition unit to acquire subgraph connection strength information, The information processing apparatus further includes a subgraph reconstruction control unit that instructs the subgraph determination unit to determine whether or not a predetermined condition is satisfied.
With such a configuration, it is possible to reconstruct a subgraph by detecting interference between nodes constituting the subgraph and detecting the interference.

Further, in the information processing apparatus according to the seventh aspect of the present invention, in contrast to the third aspect of the invention, the partial graph determination unit corresponds to one node information in the node information included in one stored partial graph information. Determining whether the connection contribution information of the first node and the subgraph connection strength information of the subgraph satisfy a predetermined condition, and the subgraph information acquisition unit is configured to determine whether the subgraph determination unit When it is determined that the above condition is satisfied, the information processing apparatus acquires partial graph information that is information related to the partial graph excluding the first node from the partial graph.
With this configuration, deletion of a node from the subgraph can be detected.

  In the information processing apparatus according to the eighth aspect of the present invention, in relation to any one of the first to seventh aspects, the partial graph information indicates node information of nodes constituting the partial graph and a link between the nodes. It has link information, and the link information is an information processing apparatus that is a directed link.

Further, in the information processing device according to the ninth aspect of the invention, with respect to any one of the first to eighth aspects, the node information is an object identifier for identifying an object, and the interaction is a conversation between objects. A content storage unit that stores at least one content including an object identifier, audio information having audio, and at least one object identifier that identifies an object facing the object identified by the object identifier. The link strength information acquisition unit detects an utterance to another object from an object identified by one object identifier based on the voice of the one or more contents, and the utterance detection unit detects Link strength information to obtain link strength information, which is information indicating the frequency of utterances The subgraph processing unit includes an acquisition unit, and when the subgraph satisfies the predetermined condition, the subgraph processing unit determines, from the content, an utterance location constituting the subgraph when the subgraph determination unit determines that the subgraph satisfies the predetermined condition. It is an information processing apparatus comprising summary extracting means for extracting and summary storing means for storing summary content that is a part of the content extracted by the summary extracting means and that has the sound of the content part.
With such a configuration, a video summary focusing on utterances can be automatically acquired.

The information processing apparatus according to the tenth aspect of the present invention is the information processing apparatus according to any one of the first to eighth aspects, wherein the node information is an object identifier for identifying an object, and the interaction is an opposite between objects. A content storage unit that stores two or more contents including an object identifier, video information having a video, and one or more object identifiers that identify an object facing the object identified by the object identifier. The link strength information acquisition unit detects an opposite of one object and another object based on one or more object identifiers that are object identifiers of the two or more contents and that identify the opposing objects. Information indicating the opposite detection means and the frequency of the opposite detection detected by the opposite detection means. Link strength information acquisition means for acquiring the link strength information, and the subgraph processing unit, when the subgraph determination unit determines that the subgraph satisfies the predetermined condition, the subgraph And a summary storage unit that stores summary content that is a part of the content extracted by the summary extraction unit and includes a video of the part of the content. Information processing apparatus.
With this configuration, it is possible to automatically obtain a video summary focusing on the opposite.

  According to the information processing apparatus of the present invention, it is possible to detect a time change of data having a graph structure composed of nodes and links.

Hereinafter, embodiments of an information processing apparatus and the like will be described with reference to the drawings. In addition, since the component which attached | subjected the same code | symbol in embodiment performs the same operation | movement, description may be abbreviate | omitted again.
(Embodiment 1)
FIG. 1 is a block diagram of an information processing apparatus according to this embodiment.

The information processing apparatus includes a node information storage unit 101, a link strength information acquisition unit 102, a subgraph connection strength information acquisition unit 103, a joint contribution degree information acquisition unit 104, a subgraph determination unit 105, a subgraph reconstruction control unit 106, a part A graph information acquisition unit 107 and a partial graph information processing unit 108 are provided.
The link strength information acquisition unit 102 includes an interaction acquisition unit 1021 and a link strength information acquisition unit 1022.

  The node information storage unit 101 stores two or more pieces of node information that are information related to the node. The node information is, for example, a node identifier that identifies the node. The node information may be information on a set of a node identifier and another node identifier linked to the node, for example. A node constitutes a graph. If the predetermined interaction is between nodes, the two nodes are connected by a link. Here, the link is usually a link having an orientation, but may be a link having no orientation. Here, the node is an abstract concept, for example, a person, data, equipment, or the like. When the node is a person, the link is information indicating an interaction such as a conversation or a handshake between people, and it is considered that the nodes are connected by the link. In addition, when the node is a person or an object (such as an exhibit or information on display), for example, the person who is the node faces the person who is the node or the object which is the node. ) Is information indicating that. Opposing is also a kind of interaction. Also, the interaction may be unidirectional (for example, an explanation given to a student by a teacher) that only affects the other, and the other may be affected. It may also act on one side (such as conversation). The node may be one data in a data group having a hypertext structure, and the link may be a link between data constituting the hypertext structure. In addition, the node is data indicating one opinion posted on the electronic bulletin board, and the link is the relationship between the data (opinions) (the relationship that responded, the relationship in favor of a certain opinion, the relationship against a certain opinion, etc.) ) May be used. Further, when the node is an electronic device, the link may be a wireless or wired network. The node information storage unit 101 is preferably a non-volatile recording medium such as a hard disk or a ROM, but can also be realized by a volatile recording medium such as a RAM.

  The link strength information acquisition unit 102 acquires node information that is information related to a node and link strength information that is information indicating the frequency of interaction between nodes at a predetermined time. The predetermined time is usually constant, but is not necessarily constant. The interaction is the above-described conversation or facing. The interaction may be data transmission between devices, jump between data in hypertext, or the like. Further, the term “opposite” does not require facing from the front, and includes a state where two nodes exist within a predetermined distance.

  In addition, the link strength information acquisition unit 102 may acquire link strength information between all nodes periodically, for example. Also, the link strength information acquisition unit 102 may periodically acquire link strength information between two nodes that do not constitute one subgraph described below, for example. In such a case, the link strength information acquisition unit 102 can acquire, for example, link strength information between nodes corresponding to node information that does not exist in the node information included in one or more stored partial graph information. The partial graph information is information related to the partial graph when the partial graph determination unit 105 determines that the partial graph satisfies a predetermined condition. The partial graph information includes, for example, node information of nodes constituting the partial graph and link information indicating links between the nodes, and the link information is a directed link. However, the link information may be information having no direction.

  The link strength information acquisition unit 102 also includes, for example, a first node corresponding to one node information in node information included in one or more stored partial graph information, and one or more stored parts. The link strength information between the second nodes corresponding to the node information that does not exist in the node information included in the graph information can be acquired. In addition, the link strength information acquisition unit 102, for example, the first node corresponding to one node information in the node information included in one stored partial graph information, and other stored partial graph information The link strength information between the second nodes corresponding to one node information in the node information possessed by can be acquired. Since the link strength information is the frequency of interaction between nodes at a predetermined time, it can be said that the link strength information is the strength of the link between nodes considering time. The calculation formula etc. which acquire link strength information are not ask | required. The data structure of link strength information does not matter. The link strength information acquisition unit 102 can be usually realized by an MPU, a memory, or the like. The processing procedure of the link strength information acquisition unit 102 is usually realized by software, and the software is recorded on a recording medium such as a ROM. However, it may be realized by hardware (dedicated circuit).

  The subgraph connection strength information acquisition unit 103 calculates the degree of connection of the subgraphs based on the one or more link strength information acquired by the link strength information acquisition unit 102, and acquires the subgraph connection strength information. A partial graph is a graph in which two or more nodes are combined. For example, the subgraph connection strength information acquisition unit 103 calculates, as the subgraph connection strength information, the sum of link strength information between all two nodes of all nodes constituting the subgraph. In addition, the subgraph connection strength information acquisition unit 103 may calculate, for example, an average value of link strength information between all two nodes of all nodes constituting the subgraph as the subgraph connection strength information. In other words, the calculation method of the partial graph connection strength information is not limited. Further, the data structure of the partial graph bond strength information is not limited. The subgraph connection strength information acquisition unit 103 can be usually realized by an MPU, a memory, or the like. The processing procedure of the subgraph connection strength information acquisition unit 103 is usually realized by software, and the software is recorded in a recording medium such as a ROM. However, it may be realized by hardware (dedicated circuit).

  The connection contribution information acquisition unit 104, based on link strength information between one node of two or more nodes constituting the subgraph and another node, the contribution of the connection in the subgraph of the one node The joint contribution, which is information indicating the degree, is calculated, and the joint contribution information is acquired. Here, the “other nodes” are all nodes except one node among the nodes constituting the subgraph. The calculation formula etc. which acquire joint contribution information are not ask | required. The data structure of the binding contribution information does not matter. The combined contribution information acquisition unit 104 can be usually realized by an MPU, a memory, or the like. The processing procedure of the combined contribution information acquisition unit 104 is usually realized by software, and the software is recorded on a recording medium such as a ROM. However, it may be realized by hardware (dedicated circuit).

  Based on the one or more link strength information acquired by the link strength information acquisition unit 102, the subgraph determination unit 105 determines whether a subgraph composed of two or more nodes and a link between the nodes is related to a predetermined degree of connectivity ( Hereinafter, it is appropriately determined whether or not “first condition” is satisfied. The first condition is, for example, a condition that the strength of coupling between two nodes is stronger than a predetermined strength of coupling. Based on the link strength information and the partial graph connection strength information, the partial graph determination unit 105 determines whether or not the partial graph satisfies a condition relating to a predetermined degree of connection (hereinafter referred to as “second condition” as appropriate). May be. The second condition is, for example, that the strength of the connection between links indicated by the link strength information is the strength of the degree of overall coupling of the subgraph indicated by the subgraph connection strength information (this subgraph has two or more nodes). Or more than the average (value for one node). Based on the link strength information and the connection contribution information, or the link strength information and the subgraph connection strength information and the connection contribution information, the subgraph determination unit 105 determines whether the subgraph has a condition related to a predetermined connection degree (hereinafter referred to as “first” It is determined whether or not “three conditions” are satisfied. The third condition is, for example, the degree of coupling indicated by the coupling contribution information, and the ratio of contribution to the coupling of the entire subgraph of one node is ½ or less of the average of the whole.

  In addition, the subgraph determination unit 105 determines whether or not the link strength information, which is information indicating the strength of the link between nodes that do not constitute any subgraph, is information indicating a degree of coupling equal to or higher than a predetermined level. As a result of such determination, when it is determined that the link strength information is information indicating a degree of coupling of a predetermined level or more, it is determined that a new subgraph has been generated.

  The subgraph determination unit 105 also includes link strength information between the first node that constitutes the subgraph, the second node that does not constitute any subgraph, and the subgraph of the subgraph having the first node. When the bond strength information satisfies a predetermined condition, it is determined that a partial graph obtained by adding a second node to the partial graph satisfies a condition regarding a predetermined degree of connection.

  In addition, the subgraph determination unit 105 has predetermined information on the link strength information between the first node constituting the first subgraph, the second node constituting the second subgraph, and the subgraph connection strength information of the first subgraph. The first node constituting the first subgraph that satisfies the condition, the link strength information between the second nodes constituting the second subgraph, and the subgraph connection strength information of the second subgraph satisfy the predetermined condition. Determine whether or not. As a result of the determination, if the predetermined condition is satisfied, the subgraph reconstruction control unit 106 described later operates in order to form a subgraph in all nodes constituting the first subgraph and the second subgraph. . That is, this determination is a determination for reconstructing the subgraph.

  In addition, the subgraph determination unit 105 is combined contribution degree information of the first node corresponding to one node information in the node information included in the stored partial graph information, and the combined contribution degree information acquisition unit It is determined whether the connection contribution information acquired by 104 and the partial graph connection strength information of the subgraph satisfy a predetermined condition. The predetermined condition is, for example, a condition for determining that the contribution degree of the connection in the subgraph of the first node is low. This determination is a determination for excluding the first node from the subgraph.

  The partial graph information is information related to the partial graph when the partial graph determination unit 105 determines that the partial graph satisfies a predetermined condition. The partial graph information includes, for example, node information of nodes constituting the partial graph and link information indicating links between the nodes, and the link information is a directed link. However, the link information may be information having no direction.

  The subgraph determination unit 105 can be usually realized by an MPU, a memory, or the like. The processing procedure of the subgraph determining unit 105 is usually realized by software, and the software is recorded on a recording medium such as a ROM. However, it may be realized by hardware (dedicated circuit).

  The subgraph reconstruction control unit 106 performs control for reconstructing the subgraph when the subgraph determination unit 105 determines that the predetermined condition is satisfied. The predetermined condition here means that the link strength information acquired by the link strength information acquisition unit 102 and the partial graph connection strength information of the first subgraph corresponding to the one partial graph information satisfy the predetermined condition, and the link strength information. The link strength information acquired by the acquisition unit 102 and the partial graph connection strength information of the second partial graph corresponding to other partial graph information satisfy a predetermined condition. Here, the link strength information acquired by the link strength information acquisition unit 102 is stored with the first node corresponding to one node information in the node information included in the stored partial graph information. This is link strength information between second nodes corresponding to one node information in node information included in other subgraph information.

  As a specific example of the control for reconstructing the subgraph, there are the following operations. For example, the subgraph reconstruction control unit 106 instructs the link strength information acquisition unit 102 to acquire link strength information between nodes constituting the first partial graph and the second partial graph, or the link strength information acquisition unit 102. Instructing to acquire link strength information between nodes constituting the first partial graph and the second partial graph and instructing the partial graph connection strength information acquisition unit to acquire the partial graph connection strength information, Instructing the subgraph determination unit to determine whether or not a predetermined condition is satisfied.

  When the partial graph determination unit 105 determines that the partial graph satisfies a predetermined condition, the partial graph information acquisition unit 107 acquires partial graph information that is information related to the partial graph. The partial graph information includes, for example, node information of nodes constituting the partial graph and link information indicating links between the nodes, and the link information is a directed link. However, the link information may be information having no direction. The subgraph information acquisition unit 107 normally accumulates the acquired subgraph information at least temporarily. Such a storage medium may be a non-volatile recording medium such as a hard disk or a ROM, or a volatile recording medium such as a RAM.

  The partial graph information processing unit 108 performs processing based on the partial graph information acquired by the partial graph information acquisition unit 107. There are various types of processing as follows. In other words, processing is processing that automatically acquires video summaries, processing that assumes topics from data posted on electronic bulletin boards (opinions), and partial data of data posted on electronic bulletin boards (opinions). These include automatic extraction processing and processing for automatically acquiring Web HP. That is, it is possible to include all processes in which it is effective to acquire a partial graph with a temporal interaction among data and devices having a graph structure. The subgraph information processing unit 108 can be realized by, for example, an MPU or a memory. The processing procedure of the subgraph information processing unit 108 is realized by software, for example, and the software is recorded on a recording medium such as a ROM. However, it may be realized by hardware (dedicated circuit). However, it goes without saying that the means for realizing the partial graph information processing unit 108 may differ depending on the processing content.

  The interaction acquisition unit 1021 acquires an interaction between two nodes. Here, the interaction is, for example, conversation or facing. The interaction may be data transmission between devices, jump between data in hypertext, or the like.

The link strength information acquisition unit 1022 acquires link strength information based on the interaction between nodes acquired by the interaction acquisition unit 1021. The link strength information is information indicating the frequency of interaction per predetermined time between nodes. The link strength information is, for example, the time (for example, 45 seconds) during which two nodes, who are nodes, have a conversation during one minute. In addition, the link strength information is, for example, the time (for example, 53 seconds) when the person who is the node was looking at the exhibition that was the node (facing the exhibition).
Next, the operation of the information processing apparatus will be described using the flowcharts of FIGS.

(Step S201) The subgraph information acquisition unit 107 substitutes 1 for a counter i. The counter i is a window width number. For example, the window width is shifted as follows. If the partial graph cannot be constructed, the time is shifted by 1 second. If the partial flag can be constructed, the time is shifted by the window width. The window width is a time width for determining whether or not a subgraph is constructed, and is, for example, 100 seconds. However, the method for shifting the window width and the length are not limited.
(Step S202) The partial graph information acquisition unit 107 acquires a partial graph in the section of the i-th window width. Details of this partial graph acquisition process will be described with reference to the flowchart of FIG.
(Step S203) The subgraph information processing unit 108 performs processing based on the subgraph information of the subgraph acquired in step S202. A specific example of such processing will be described in Embodiment 2.

(Step S204) The subgraph information acquisition unit 107 increments the counter i by 1. That is, the start point (time t) and end point (time t) indicating the next window width are determined here. An example of the process for such determination is the above-described process for shifting the window width.
(Step S205) The subgraph determining unit 105 detects a change in the subgraph. The details of this change detection process will be described with reference to the flowchart of FIG.

  (Step S206) The subgraph information acquisition unit 107 determines whether or not a change in the subgraph is detected in step S205. If a change in the subgraph is detected, the process proceeds to step S207, and if a change in the subgraph is not detected, the process jumps to step S209.

(Step S207) The partial graph information acquisition unit 107 updates the partial graph. The update of the partial graph is a process of updating the partial graph information based on the change of the partial graph. Details of the update process of the subgraph will be described with reference to the flowchart of FIG.
(Step S208) The subgraph information processing unit 108 performs processing based on the subgraph information of the subgraph updated in step S207. A specific example of such processing will be described in Embodiment 2.
(Step S209) The subgraph information acquisition unit 107 increments the counter i by 1.

(Step S210) The subgraph information acquisition unit 107 determines whether or not to end the process. Specifically, the subgraph information acquisition unit 107 determines whether there is information on temporal transition of the graph having the i-th window width. If it is determined that the process is not terminated, the process returns to step S205. If it is determined that the process is to be terminated, the process is terminated.
Next, the operation of the partial graph acquisition process will be described using the flowchart of FIG.
(Step S301) The link strength information acquisition unit 102 substitutes 1 for a counter i.

  (Step S <b> 302) The link strength information acquisition unit 102 determines whether there is an i-th node set. If the i-th node set exists, the process proceeds to step S303. If the i-th node set does not exist, the process jumps to step S307. Specifically, for example, the link strength information acquisition unit 102 determines whether or not the i-th node set exists by the following processing. The node information storage unit 101 stores node identifiers of all nodes (here, for example, the node identifier is an object identifier). Then, the link strength information acquisition unit 102 sequentially determines combinations of two node identifiers in all node identifiers, and acquires information on the combinations. In such a case, the link strength information acquisition unit 102 determines whether or not there is a set of i-th nodes, depending on whether or not all the combinations of information have been acquired. That is, when the node information storage unit 101 stores three node identifiers, the maximum value of i is “3”.

  (Step S303) The interaction acquisition unit 1021 acquires information regarding the interaction in the window width between the i-th node pair. The interaction acquisition unit 1021 acquires, for example, an utterance section from the user “A” to the user “B” and an utterance section from the user “B” to the user “A” in 100 seconds (window width). Specifically, for example, the link strength information acquisition unit 102 analyzes the acquired voice, determines that the voice of a predetermined level or more is being spoken, and performs processing such as acquiring the utterance section. .

  (Step S304) The link strength information acquisition unit 1022 calculates the frequency of interaction (for example, utterance) per predetermined time (so-called window width) from one node constituting the i-th node set to the other node (here, utterance). Then, link strength information that is time is acquired. Specifically, the link strength information acquisition unit 1022 acquires the utterance time (for example, 60 seconds) from the user “A” to the user “B”.

  (Step S305) The link strength information acquisition unit 1022 acquires link strength information that is the frequency of interaction per predetermined time from the other node constituting the i-th node set to the one node. Specifically, the link strength information acquisition unit 1022 acquires the utterance time (for example, 50 seconds) from the user “B” to the user “A”.

(Step S306) The link strength information acquisition unit 102 increments the counter i by one. The process returns to step S302. Note that the link strength information between all nodes has been acquired by the processing from step S301 to step S306.
(Step S307) The subgraph determining unit 105 substitutes 0 for the counter i.

(Step S308) The subgraph determination unit 105 acquires the total number of nodes from the node information storage unit 101. Note that the total number of nodes may be acquired, for example, when the interaction acquisition unit 1021 detects an interaction. In such a case, the information processing apparatus may not have the node information storage unit 101.
(Step S309) The subgraph determining unit 105 substitutes 1 for a counter j.

(Step S310) The subgraph determining unit 105 determines whether there is a combination of jth (N−i) nodes. If the j-th (Ni) node combination exists, the process goes to step S311. If the j-th (Ni) node combination does not exist, the process jumps to step S312.
(Step S311) The subgraph determining unit 105 acquires j-th (N−i) node information. Go to step S314.
(Step S312) The subgraph determination unit 105 increments the counter i by one.
(Step S313) The subgraph determining unit 105 determines whether (N−i) is 2 or more. If (Ni) is 2 or more, the process returns to step S309, and if (Ni) is not 2 or more, the process returns to the upper function.

(Step S314) The subgraph determining unit 105 selects all the nodes of the j-th (N−i) node information acquired in step S311 (referred to as “node information group” as appropriate since it has a plurality of node information). It is determined whether or not the included partial graph exists in the partial graph corresponding to the partial graph information already acquired in step S317. If such a subgraph does not exist, the process goes to step S315, and if it exists, the process jumps to step S318.
(Step S315) The subgraph determination unit 105 acquires partial graph information of this node information group, and acquires all link strength information between all two nodes constituting the partial graph information.

  (Step S316) Based on the total link strength information acquired in step S315, the subgraph determination unit 105 determines whether a subgraph composed of two or more nodes and a link between the nodes satisfies a condition regarding a predetermined degree of connectivity. Judge whether or not. If the condition is satisfied, the process goes to Step S317, and if the condition is not satisfied, the process jumps to Step S318. A specific example of the predetermined condition will be described later. The concept of the predetermined condition will be described as follows. That is, it is whether or not the degree of connection of the subgraph composed of links between two or more nodes is greater than the predetermined degree of connection. If the degree of connection is greater than a predetermined degree, two or more extracted nodes in a predetermined window width constitute a subgraph.

(Step S317) The partial graph information acquisition unit 107 acquires partial graph information corresponding to the nodes constituting this node information group. Such partial graph information may be a combination of information of only a set of node identifiers and information of time (start point, end point) constituting a window width, or information of time constituting a set of node identifiers and window width and total link strength. Information may be used.
(Step S318) The subgraph determining unit 105 increments the counter j by 1. The process returns to step S310.
Next, the operation of the change detection process of the information processing apparatus will be described using the flowchart of FIG.
(Step S401) The interaction acquisition unit 1021 acquires all the interactions between nodes at a predetermined time point.

  (Step S402) The subgraph determining unit 105 determines whether or not an interaction is acquired in step S401. If the interaction is acquired, the process goes to step S403, and if the interaction is not acquired, the process returns to the upper function. If no interaction is acquired, no change has been detected in the change detection process.

(Step S403) The subgraph determining unit 105 performs a subgraph generation determination that is a determination as to whether or not a new subgraph can be generated. The process of determining the partial graph will be described with reference to the flowchart of FIG.
(Step S404) The subgraph determination unit 105 goes to step S412 if a subgraph can be generated, and goes to step S405 if a subgraph cannot be generated.

(Step S <b> 405) The subgraph determination unit 105 performs node addition determination that is a determination as to whether or not a new node can be added to the subgraph. The node addition determination process will be described with reference to the flowchart of FIG.
(Step S406) The subgraph determination unit 105 goes to step S412 if a node can be added, and goes to step S407 if a node cannot be added.

(Step S407) The subgraph determining unit 105 determines whether or not an interaction exists between a node constituting one subgraph and a node constituting another subgraph. Such an interaction is called interference between subgraphs. Such subgraph interference determination processing will be described with reference to the flowchart of FIG.
(Step S408) The subgraph determination unit 105 goes to step S411 if there is interference between subgraphs, and goes to step S409 if there is no interference between subgraphs.

  (Step S409) The subgraph determining unit 105 determines whether or not to delete a node from one subgraph. Such a determination is called a node deletion determination. The node deletion determination process will be described with reference to the flowchart of FIG.

(Step S410) The subgraph determining unit 105 determines whether or not there is a subgraph from which a node is deleted as a result of the determination in step S409. If there is a subgraph for deleting a node, the process goes to step S411, and if it does not exist, the process returns to the upper function.
(Step S411) The subgraph information acquisition unit 107 performs a process of acquiring subgraph information between all nodes constituting two subgraphs with interference. Such processing is the processing of the flowchart of FIG.

(Step S412) The partial graph information including the node information of the newly generated partial graph, the partial graph to which the node is added, or the partial graph from which the node is deleted is configured and acquired. Return to upper function. In such a case, a return value indicating that a change has been detected is returned to the upper function.
In the flowchart of FIG. 4, it is assumed that the above process is usually repeated every time an interaction is detected.
Next, the operation of the partial graph generation determination of the information processing apparatus will be described using the flowchart of FIG.
(Step S501) The subgraph determination unit 105 substitutes 1 for a counter i. Here, the two nodes to be interacted with are referred to as one node and the other node.

  (Step S502) The subgraph determining unit 105 determines whether or not the i-th subgraph information exists. The i-th subgraph information is subgraph information currently stored at least temporarily (subgraph information currently stored as constituting a subgraph, and information on subgraphs that have already disappeared is Is the i-th subgraph information. If the i-th subgraph information exists, the process goes to step S503, and if the i-th subgraph information does not exist, the process jumps to step S507.

  (Step S503) The subgraph determining unit 105 determines whether node information of one node exists in the i-th subgraph information. If the node information of one node exists, the process goes to step S506, and if the node information of one node does not exist, the process goes to step S504.

(Step S504) The subgraph determining unit 105 determines whether or not the node information of the other node exists in the i-th subgraph information. If the node information of the other node exists, the process goes to step S506, and if the node information of the other node does not exist, the process goes to step S505.
(Step S505) The subgraph determining unit 105 increments the counter i by 1. The process returns to step S502.
(Step S <b> 506) The subgraph determination unit 105 substitutes information indicating that a subgraph is not generated for the return value. Return to upper function.

  (Step S507) The link strength information acquisition unit 102 acquires link strength information between two nodes to be interacted with. When the link strength information acquisition unit 102 has already acquired link strength information between two nodes and temporarily stores the link strength information, the subgraph determination unit 105 reads the link strength information.

(Step S508) The subgraph determining unit 105 determines whether or not the link strength information acquired in step S507 is information indicating a degree of coupling equal to or higher than a predetermined level. If the information indicates a degree of coupling greater than or equal to a predetermined level, the process proceeds to step S509.
(Step S509) The subgraph determining unit 105 substitutes information indicating that a subgraph is to be generated into the return value. Return to upper function.
Next, the node addition determination operation of the information processing apparatus will be described with reference to the flowchart of FIG.
(Step S601) The subgraph determination unit 105 substitutes 1 for a counter i.

  (Step S602) The subgraph determining unit 105 determines whether or not the i-th subgraph information exists. The i-th subgraph information is i-th subgraph information among the subgraph information currently stored at least temporarily. If the i-th subgraph information exists, the process goes to step S603, and if the i-th subgraph information does not exist, the process jumps to step S610.

  (Step S603) The subgraph determining unit 105 determines whether node information of one node exists in the i-th subgraph information. If node information of one node exists, the process goes to step S606, and if node information of one node does not exist, the process goes to step S612.

  (Step S604) The subgraph determining unit 105 determines whether the other node belongs to any of the subgraphs. Such determination processing is determined as follows. It is checked whether or not the subgraph information having the node information of the other node exists in the subgraph information stored at least temporarily.

  (Step S605) The subgraph determination unit 105 determines whether the other node belongs to any subgraph or does not belong to any subgraph as a result of the inspection in step S604. If it does not belong to any subgraph, it goes to step S606, and if it belongs to any subgraph, it goes to step S612.

(Step S606) The link strength information acquisition unit 102 acquires link strength information between two nodes to be interacted with. When the link strength information acquisition unit 102 has already acquired link strength information between two nodes and temporarily stores the link strength information, the subgraph determination unit 105 reads the link strength information.
(Step S607) The subgraph connection strength information acquisition unit 103 acquires subgraph connection strength information corresponding to the i-th subgraph information.

(Step S608) The subgraph determining unit 105 determines whether or not the link strength information acquired in step S606 and the subgraph connection strength information acquired in step S607 satisfy a predetermined condition. If the predetermined condition is satisfied, the process goes to step S609, and if the predetermined condition is not satisfied, the process jumps to step S612. The concept of the predetermined condition is as follows. That is, the predetermined condition is that the degree of connection of the added node with the subgraph is greater than the predetermined degree.
(Step S609) The subgraph determining unit 105 configures a return value for adding a node. Return to upper function.
(Step S610) The subgraph determining unit 105 configures a return value indicating that a node is not added. Return to upper function.

(Step S611) The subgraph determining unit 105 determines whether or not the node information of the other node exists in the i-th subgraph information. If the node information of the other node exists, the process goes to step S604, and if the node information of the other node does not exist, the process goes to step S612.
(Step S612) The subgraph determining unit 105 increments the counter i by 1. The process returns to step S602.
Next, the operation of subgraph interference determination of the information processing apparatus will be described using the flowchart of FIG.
(Step S701) The subgraph determination unit 105 substitutes 1 for a counter i. Here, the two nodes to be interacted with are referred to as one node and the other node.

  (Step S702) The subgraph determining unit 105 determines whether or not the i-th subgraph information exists. The i-th subgraph information is i-th subgraph information among the subgraph information currently stored at least temporarily. If the i-th subgraph information exists, the process proceeds to step S703, and if the i-th subgraph information does not exist, the process jumps to step S712.

(Step S703) The subgraph determining unit 105 determines whether node information of one node exists in the i-th subgraph information. If node information of one node exists, the process goes to step S704, and if node information of one node does not exist, the process goes to step S713.
(Step S704) The subgraph determination unit 105 substitutes “i + 1” for the counter j. Here, it is assumed that the partial graph information is stored in order.

  (Step S705) The subgraph determining unit 105 determines whether or not the j-th subgraph information exists. If the jth subgraph information exists, the process goes to step S706, and if the jth subgraph information does not exist, the process goes to step S713.

(Step S706) The subgraph determining unit 105 determines whether or not the other node information exists in the j-th subgraph information. If the other node information exists, the process goes to step S707, and if the other node information does not exist, the process goes to step S714.
(Step S707) The subgraph connection strength information acquisition unit 103 acquires partial graph connection strength information of the subgraph indicated by the i-th subgraph information.
(Step S708) The subgraph connection strength information acquisition unit 103 acquires the subgraph connection strength information of the subgraph indicated by the j-th subgraph information.
(Step S709) The link strength information acquisition unit 102 acquires link strength information between two nodes that have interacted with each other.

(Step S710) The subgraph determining unit 105 determines whether or not the two subgraph connection strength information acquired in steps S707 and S708 and the link strength information acquired in step S709 satisfy a predetermined condition. If the predetermined condition is satisfied, the process proceeds to step S711. If the predetermined condition is not satisfied, the process proceeds to step S713.
(Step S711) The subgraph determining unit 105 substitutes a value indicating that there is inter-subgraph interference in the return value. Return to upper function.
(Step S712) The subgraph determining unit 105 substitutes a value indicating that there is no intergraph interference as a return value. Return to upper function.
(Step S713) The subgraph determining unit 105 increments the counter i by 1. Go to step S702.
(Step S714) The subgraph determining unit 105 increments the counter j by 1. Go to step S705.
Next, the node deletion determination operation of the information processing apparatus will be described with reference to the flowchart of FIG.
(Step S801) The subgraph judgment unit 105 substitutes 1 for a counter i. Here, the two nodes to be interacted with are referred to as one node and the other node.

(Step S802) The subgraph determining unit 105 determines whether or not the i-th subgraph information exists. If the i-th subgraph information exists, the process goes to step S503, and if the i-th subgraph information does not exist, the process returns to the upper function.
(Step S803) The subgraph connection strength information acquisition unit 103 acquires partial graph connection strength information of a subgraph corresponding to the i-th subgraph information.
(Step S804) The subgraph determining unit 105 substitutes 1 for a counter j.

(Step S805) The subgraph determining unit 105 determines whether or not the j-th node exists in the subgraph corresponding to the i-th subgraph information. If the jth node exists, the process goes to step S806, and if the jth node does not exist, the process goes to step S814.
(Step S806) The subgraph determination unit 105 substitutes 1 for a counter k.
(Step S807) The subgraph determining unit 105 determines whether or not the kth node excluding the jth node exists.
(Step S808) The link strength information acquisition unit 102 acquires link strength information between the j-th node and the k-th node.
(Step S809) The subgraph determining unit 105 increments the counter k by 1. The process returns to step S807.
(Step S810) The combined contribution information acquisition unit 104 acquires combined contribution information of the j-th node. Such joint contribution information is acquired based on the one or more link strength information acquired in step S808.

  (Step S811) The subgraph determining unit 105 determines whether or not the combined contribution information acquired in step S810 satisfies a predetermined condition. This condition is a condition indicating that the j-th node has a higher contribution than the predetermined contribution to the combination of subgraphs corresponding to the i-th subgraph information. If the condition is satisfied, the process goes to step S812. If the condition is not satisfied, the process goes to step S813.

(Step S812) The subgraph information acquisition unit 107 performs a process for deleting the jth node from the subgraph corresponding to the i th subgraph information. Such processing is, for example, processing for deleting new node information corresponding to the j-th node from the i-th subgraph information to form new subgraph information.
(Step S813) The subgraph determining unit 105 increments the counter j by 1. Go to step S805.
(Step S814) The subgraph determining unit 105 increments the counter i by 1. Go to step S802.
Next, the operation of the partial graph information update process of the information processing apparatus will be described with reference to the flowchart of FIG.

(Step S901) The partial graph information acquisition unit 107 determines whether or not there is a change that a partial graph is generated. If there is a change to generate a subgraph, the process goes to step S902, and if there is no change, the process goes to step S904.
(Step S902) The subgraph information acquisition unit 107 configures subgraph information of a subgraph to be generated based on node information of two nodes that have interacted with each other.

  (Step S903) The partial graph information acquisition unit 107 adds the partial graph information configured in step S902. Such additional writing is performed, for example, on a partial graph information management table (details will be described later). Return to upper function.

(Step S904) The subgraph information acquisition unit 107 determines whether or not there is a change that a node is added to the subgraph. If there is a change that a node is added to the subgraph, the process goes to step S905, and if there is no change, the process goes to step S907.
(Step S905) The partial graph information acquisition unit 107 configures partial graph information to which node information of the node to be added is added.
(Step S906) The partial graph information acquisition unit 107 overwrites and updates the partial graph information configured in step S905 with the partial graph information of the original partial graph. Return to upper function.

(Step S907) The subgraph information acquisition unit 107 determines whether there is interference between subgraphs. If there is interference between subgraphs, the process goes to step S908, and if there is no interference between subgraphs, the process goes to step S911.
(Step S908) The partial graph information acquisition unit 107 deletes the partial graph information of the original two partial graphs. Such deletion is, for example, deletion from the subgraph information management table.

(Step S909) The partial graph information acquisition unit 107 determines whether or not the generated partial graph information exists. If the generated partial graph information exists, the process goes to step S910, and if the generated partial graph information does not exist, the process returns to the upper function.
(Step S910) The subgraph information acquisition unit 107 adds all the subgraph information to, for example, a subgraph information management table. Return to upper function.
(Step S911) The subgraph information acquisition unit 107 determines whether there is a node deletion. If there is a node deletion, the process goes to step S912. If there is no node deletion, the process returns to the upper function.
(Step S912) The subgraph information acquisition unit 107 configures the subgraph information from which the node information has been deleted, or reads out the subgraph information from which the already configured node information has been deleted.
(Step S913) The partial graph information acquisition unit 107 overwrites and updates the partial graph information acquired in step S912. Return to upper function.

In the flowchart of FIG. 9, the newly acquired partial graph information may be added without being updated. In such a case, it is desirable that the partial graph information includes information on a time zone in which the partial graph is valid (for example, information on a start point and an end point described later).
Needless to say, the various types of processing described in the above flowcharts are examples.
Hereinafter, the concept of the operation of the information processing apparatus in the present embodiment will be described.

  The information processing apparatus performs processing for extracting a subgraph that is a portion of the graph from the graph structure formed by the nodes and the links based on the weights given to the links. In particular, the information processing apparatus introduces the concept of time, and performs processing for extracting a coherent subgraph of a coherent time based on the degree of coupling between nodes that changes with the passage of time. Do. Further, in the information processing apparatus, since the objects in the space are captured as nodes constituting the graph, the graph structure is arranged in the space and moved in the space.

Here, the node is an object such as a person or a thing. Links are created by interactions between objects. The interaction is, for example, conversation or facing. That is, when two people are talking, a link is established between the two people, and the frequency of the conversation (the conversation time of two people per unit time) indicates the link strength. The link strength is indicated by link strength information indicating the strength of the link.
In the above situation, the concept of processing in which the information processing apparatus extracts a subgraph by the operation described with reference to FIGS. 2 to 9 will be described.
First, the basic concept in the processing of the information processing apparatus, that is, the link strength, the strength of the partial graph, and the degree of contribution to the partial graph will be described with reference to FIG.

First, FIG. 10A is a conceptual diagram illustrating the strength of a link. In FIG. 10A, node “B” interacts with node “A”. Specifically, the person “B” is speaking to the person “A”. In such a case, the link strength information is calculated by, for example, Equation 1 below. That is, the link strength information is a time during which interaction (for example, speech) is performed per unit time (for example, 10 minutes). Note that the interaction may be a one-sided action on the other, such as speech or explanation, or a conversation that interacts with each other.

In Formula 1, L _AB (t) is link strength information. I _AB (t ₁ , t ₂ ) is the total time of interaction (for example, utterance) performed by A on B in the section from time t ₁ to time t ₂ . w ₁ is the window width. The window width is a unit time for determining the strength of interaction.

FIG. 10B is a conceptual diagram illustrating the strength of the entire subgraph having two or more nodes. In FIG. 10B, the nodes “A”, “B”, “C”, and “D” interact with each other. Specifically, each person (“A”, “B”, “C”, “D”) is in a state of discussing by talking to other people. In such a case, the subgraph bond strength information is calculated by the following formula 2, for example. That is, the subgraph connection strength information is calculated using all link strength information between two nodes as parameters. More specifically, for example, the subgraph connection strength information is a value obtained by dividing the sum of all link strength information between two nodes by the number of nodes.

In Equation 2, C _x (t) is the subgraph connection strength information, and N _x is the number of nodes constituting the subgraph. T _x (t) (i, j) is obtained by changing each element to the strength of connection between nodes with respect to the adjacency matrix of the graph structure representing the subgraph X, and the value of the i, j component is Denote as T _x (t) (i, j). Also, t _s indicates a time at which the subgraph is created. In addition, _{w 2} is the window width.

FIG. 10C is a conceptual diagram for explaining the joint contribution, which is information indicating the joint contribution in the subgraph of one node. In FIG. 10C, the node “A” is acting on the other person, or the node “A” is acting on the other person. Information indicating the degree of action of the node “A” on the subgraph when viewed from the node “A” is the combined contribution information. The joint contribution information is calculated by, for example, Equation 3 below.

In Equation 3, CO _sx (t) is joint contribution information. Also. N _X is the number of nodes existing in the subgraph X, U _X (t) is an adjacency matrix of the graph structure representing the subgraph X, and each element contributes to the subgraph X It is changed to the strength of link ties. In addition, _{w 3} it is the window width.
The information processing apparatus configures a partial graph using any one of the above link strength information, partial graph connection strength information, and connection contribution information, or information on a combination thereof.
Next, the concept of processing when the information processing apparatus generates a partial graph will be described with reference to the conceptual diagram of FIG.

The information processing apparatus acquires link strength information between the node “A” and the node “B”. It is assumed that node “A” and node “B” do not belong to any subgraph. Then, when the link strength information between the node “A” and the node “B” is link strength information that is equal to or greater than predetermined link strength information (for example, 1 minute), the node “A” and the node “B” ”As node information. Then, at least temporarily, the information processing apparatus accumulates partial graph information that is a record of “ID = 1” illustrated in FIG. FIG. 12 is a partial graph information management table storing partial graph information. The partial graph information management table holds one or more records having “ID” and “partial graph information”. “ID” is information for identifying a record and exists for record management in the table. The “partial graph information” includes “node information 1”, “node information 2”, and “link strength information”. “Node information 1” and “node information 2” are node information, and here are identifiers for identifying the nodes. “Partial graph information” indicates that an action is applied from the node of node information 1 to the node of node information 2 and is connected by a directed link. The “link strength information” is a value indicating the strength of the link from the node of the node information 1 to the node of the node information 2. Here, it is assumed that a partial graph is formed when the link strength information is a value of 1 or more. In other words, if the node is a person and the interaction is an utterance, the information processing apparatus recognizes that the state of talking for more than 1 minute is the state constituting the subgraph. In the record of “ID = 1” in FIG. 12, the user “A” speaks to the user “B” for 2.5 minutes, and the user “B” contacts the user “A” for 1.8 minutes. Indicates that you have spoken. A specific example of how to determine the state of talking for 1 minute or more will be described later.
Next, the concept of node addition processing, which is processing for adding a node to a subgraph and forming a new subgraph, will be described with reference to the conceptual diagram of FIG.

  In FIG. 13, a partial graph having nodes “A”, “B”, and “C” is configured, and a node “D” is added to the graph. In such a case, the information processing apparatus detects the interaction between the node “D” and the node “C”, and calculates link strength information based on the interaction. Then, the information processing apparatus calculates partial graph connection strength information of the partial graph having the nodes “A”, “B”, and “C”. Then, the information processing apparatus determines whether or not to add the node “D” to the subgraphs of the nodes “A”, “B”, and “C” based on the calculated link strength information and the calculated subgraph connection strength information. decide. Here, for example, when the calculated link strength information is at least twice as large as the calculated subgraph connection strength information, the information processing apparatus determines to add the node “D” to the subgraph, for example. To do. In FIG. 13, “>>” means that it is sufficiently large, and the information processing apparatus previously holds a value that can be determined to be sufficiently large. Here, the original value for determining that it is sufficiently large is, for example, “double”.

  Next, the concept of interference processing between subgraphs, which is processing for reconstructing a subgraph by interference between subgraphs, will be described with reference to the conceptual diagram of FIG. In FIG. 14, there are a first subgraph having nodes “A”, “B”, and “C” and a second subgraph having nodes “D” and “E”. That is, at a certain time, the first subgraph information based on the first subgraph having the nodes “A”, “B”, and “C”, and the second subgraph based on the second subgraph having the nodes “D” and “E”. Information is generated and held by the information processing apparatus. In this situation, it is assumed that the information processing apparatus has detected an interaction from the node “D” to the node “C”. The information processing apparatus calculates link strength information between the node “D” and the node “C” based on the interaction. In addition, the information processing apparatus calculates partial graph connection strength information of the first partial graph. Further, the information processing apparatus calculates partial graph connection strength information of the second partial graph. Then, the information processing apparatus determines whether the calculated link strength information has a predetermined relationship with the partial graph connection strength information of the first partial graph and the partial graph connection strength information of the second partial graph. Here, the predetermined relationship is, for example, that the value indicated by the link strength information is sufficiently larger than the value indicated by the partial graph connection strength information of the first subgraph, and the value indicated by the link strength information is the subgraph connection of the second subgraph. It is sufficiently larger than the value indicated by the intensity information. Note that “sufficiently large” means, for example, that the difference is greater than or equal to a predetermined value difference as described above. When determining that the information processing apparatus has the above-described predetermined relationship, the information processing apparatus determines that there is a possibility that a new subgraph is configured, and nodes “A”, “B”, “C”, “D”, and “E”. The process returns to the first step to determine whether or not a subgraph exists, and detects a new subgraph. In such a case, the subgraph is reconfigured because the node “D” has left the second subgraph and is connected to the first subgraph, or the node “C” has left the first subgraph and is in the second subgraph. This is because it cannot be determined at this point whether the nodes “C” and “D” have joined each subgraph to form a new subgraph. Therefore, if a process of detecting a new subgraph in the nodes “A”, “B”, “C”, “D”, and “E” is performed, a new subgraph at this time can be accurately detected.

  Next, the concept of the node deletion process that is a process of deleting one or more nodes from the partial graph will be described with reference to the conceptual diagram of FIG. In FIG. 15, there is a subgraph having nodes “A”, “B”, “C”, and “D”. That is, the information processing apparatus holds partial graph information based on the partial graph having the nodes “A”, “B”, “C”, and “D” at a certain point in time. Then, for example, the information processing apparatus periodically checks the stored partial graph information, and determines whether there is a node that should be removed from the partial graph by the above-described processing. Specifically, for example, the information processing apparatus calculates joint contribution information of the node “D” in the subgraph of FIG. If the calculated value of the connection contribution information is sufficiently smaller than the value of the partial graph connection strength information of the subgraph, the information processing apparatus should delete the node “D” from the subgraph. Judge that there is. Then, new subgraph information (subgraph information from which the node “D” is deleted) is configured. Then, the information processing apparatus overwrites and updates the new partial graph information over the original partial graph information.

  The above is an example of a segment graph segmentation algorithm. The node group constituting the subgraph changes with time, but this information processing apparatus detects such a change and acquires a series of subgraph information until the subgraph information acquisition unit changes, The partial graph information processing unit performs various processes based on the partial graph information.

  As described above, according to the present embodiment, it is possible to detect temporal changes in data having a graph structure composed of nodes and links, detect temporally changing subgraphs (graphs formed by combining nodes), and A set of subgraph information can be acquired. And if a predetermined process is performed based on the acquired partial graph information, information processing with various effects becomes possible. A specific example of such information processing will be described later.

  Note that the processing in the present embodiment may be realized by software. Then, this software may be distributed by software download or the like. Further, this software may be recorded and distributed on a recording medium such as a CD-ROM. This also applies to other embodiments in this specification. Note that the software that implements the information processing apparatus according to the present embodiment is the following program. That is, the program stores, in the computer, link strength information acquisition step for acquiring node information that is information about the node and link strength information that is information indicating the frequency of interaction between the nodes per predetermined time, and the link strength. Subgraph determination that determines whether or not a subgraph composed of two or more nodes and links between the nodes satisfies a condition related to a predetermined degree of connection based on one or more link strength information acquired in the information acquisition step A subgraph information acquisition step of acquiring subgraph information that is information related to the subgraph when the subgraph is determined to satisfy the predetermined condition in the step and the subgraph determination step; and the subgraph information A program for executing a subgraph information processing step that performs processing based on That.

  Further, in the above program, the subgraph connection strength information for calculating the degree of connection of the subgraphs based on the one or more link strength information acquired in the link strength information acquisition step and acquiring the subgraph connection strength information. A program for further executing an acquisition step, and determining whether or not the partial graph satisfies a condition related to a predetermined degree of coupling based on the link strength information and the partial graph connection strength information in the partial graph determination step; But it ’s okay.

  In the above program, based on link strength information between one node of two or more nodes constituting the subgraph and another node, the degree of contribution of the connection of the one node in the subgraph is calculated. A connection contribution information acquisition step of calculating a connection contribution that is information to be acquired and acquiring connection contribution information is further executed, and in the subgraph determination step, the link strength information and the connection contribution information, or the link A program for determining whether or not the partial graph satisfies a condition relating to a predetermined degree of coupling based on strength information, the partial graph coupling strength information, and the coupling contribution degree information may be used.

  In the above program, in the link strength information acquisition step, link strength information between nodes corresponding to node information that does not exist in the node information included in the one or more stored partial graph information is acquired. In the graph determining step, a program for determining whether or not the link strength information acquired in the link strength information acquiring step is information indicating a degree of coupling equal to or higher than a predetermined value may be used.

  In the program, in the link strength information acquisition step, the first node corresponding to one node information in the node information included in the one or more stored partial graph information, and the stored 1 The link strength information between the second nodes corresponding to the node information that does not exist in the node information included in the subgraph information is acquired, and the link acquired in the link strength information acquisition step in the subgraph determination step When the strength information and the subgraph connection strength information of the subgraph having the first node satisfy a predetermined condition, the subgraph in which the second node is added to the subgraph has a condition regarding a predetermined degree of connectivity. It may be a program that is judged to satisfy.

  In the above program, in the link strength information acquisition step, the first node corresponding to one node information in the node information included in the stored partial graph information, and the other stored Link strength information between the second nodes corresponding to one node information in the node information included in the subgraph information is acquired, and the link strength information acquired in the link strength information acquisition step in the subgraph determination step And the first subgraph corresponding to the one subgraph information satisfies a predetermined condition, and the link strength information acquired in the link strength information acquisition step and the second subgraph information corresponding to the other subgraph information. It is determined whether or not the subgraph connection strength information of the two subgraphs satisfies a predetermined condition, and the predetermined graph is determined by the subgraph determination. When it is determined that the condition is satisfied, an instruction is given to acquire link strength information between nodes constituting the first partial graph and the second partial graph, or the first partial graph and the second partial graph are configured. Subgraph reconstruction control instructing to acquire link strength information between nodes, instructing to acquire subgraph connection strength information, and then instructing whether or not a predetermined condition is satisfied It may be a program that further causes the computer to execute the steps.

Further, in the above program, in the subgraph determination step, the connection contribution degree information of the first node corresponding to one node information in the node information included in the stored subgraph information and the subgraph It is determined whether or not the subgraph connection strength information satisfies a predetermined condition, and in the subgraph information acquisition step, when it is determined that the predetermined condition is satisfied in the subgraph determination step, the subgraph is obtained from the subgraph. A program that acquires subgraph information, which is information about the subgraph excluding the first node, may be used.
(Embodiment 2)

  In this embodiment, a specific example of processing in the information processing apparatus described in Embodiment 1 will be described. The information processing apparatus is an apparatus that automatically creates a video summary. In the present embodiment, the node information is an object identifier that identifies an object, and the interaction is a conversation between objects. Here, the object is a person. The object may be an exhibit or a display panel. The interaction here is opposite to speech. Speaking is talking from one person to another. Opposing means that one person faces the other person or object (exhibit or the like).

  FIG. 16 is a block diagram of the information processing apparatus in this embodiment. The information processing apparatus includes a content storage unit 1601, a link strength information acquisition unit 1602, a subgraph connection strength information acquisition unit 103, a joint contribution information acquisition unit 104, a subgraph determination unit 105, a subgraph reconstruction control unit 106, a part A graph information acquisition unit 107 and a partial graph information processing unit 1608 are provided.

  The link strength information acquisition unit 1602 includes an utterance detection unit 16021, an opposite detection unit 16022, and a link strength information acquisition unit 16023. The subgraph information processing unit 1608 includes summary extraction means 16081 and summary storage means 16082.

  The content storage unit 1601 stores one or more contents. The content includes an object identifier, video information including video, audio information including audio, and one or more object identifiers that identify an object facing the object identified by the object identifier.

  In addition, the acquisition method of the content of the content storage part 1601 does not ask | require. For example, the content is acquired by the following information acquisition device shown in FIG. In FIG. 17, the information acquisition apparatus includes a “CCD camera”, an “infrared ID tag”, an “infrared sensor”, an “IR tracker”, a “microphone”, a “throat mic”, and an “HMD (head-mounted display)”. In addition, a “CCD camera”, “infrared ID tag”, and “infrared sensor” are provided above the user's ear. The “CCD camera” acquires an image. The “infrared ID tag” transmits an infrared signal on which a signal indicating the object identifier of the user is superimposed. The “infrared sensor” receives an infrared signal from the outside. That is, the “infrared sensor” is the “IR tracker” described above. It has a “microphone” at the mouth and a “throat microphone” at the throat. In addition, an HMD (head mounted display) is provided in front of the eyes. The signal of the “CCD camera” is acquired by the PC carrying the back of the user and transmitted from the PC to the information processing apparatus. Further, the HMD is used to output information on the position of the user, the displayed exhibits and the people (objects) facing the user. Needless to say, the information acquisition apparatus shown in FIG. 17 is an example of an information acquisition method.

  Here, the situation of this example will be described. Here, two or more users wearing the information acquisition device are at a conference hall where many display panels exist, for example. In this exhibition panel, a panel to be announced (information on technical explanation) is displayed on a display having a touch panel. There are many visitors (users) of the display panel at the venue. Each user wears the information acquisition device shown in FIG. 17 and sees the video he / she sees, the object identifier for identifying the object (person, display panel, etc.) facing him, the voice uttered (a set of such information is content) Get). Then, the content is sent to the information processing apparatus.

The object identifier is, for example, an ID assigned to each visitor, and is transmitted from an infrared ID tag of an information acquisition device (see FIG. 17) attached to each visitor. The object identifier is included in the content transmitted from the information acquisition device worn by each visitor to the information processing device. That is, the video acquired by the CCD camera, the audio acquired by the microphone, the object identifier acquired by the IR tracker (identifier of the person facing the visitor), and the like are associated by the object identifier that identifies the visitor. Note that the IR tracker of the information acquisition device worn by each visitor obtains an ID (object identifier) transmitted from the infrared ID tag of the information acquisition device worn by the opposite person.
The content storage unit 1601 is preferably a non-volatile recording medium such as a hard disk or ROM, but can also be realized by a volatile recording medium such as a RAM.

  The utterance detection unit 16021 detects an utterance for another object from the object identified by one object identifier based on the sound of one or more contents. It is preferable that the utterance detection unit 16021 detects an utterance to another object from the object identified by one object identifier based on the sound of one or more contents and the object identifier identifying the opposite object. . A specific speech detection algorithm of the speech detection means 16021 will be described later. The utterance detection unit 16021 can usually be realized by an MPU, a memory, or the like. The processing procedure of the speech detection means is usually realized by software, and the software is recorded on a recording medium such as a ROM. However, it may be realized by hardware (dedicated circuit).

  Opposite detection means 16022 is an object identifier that two or more contents have, and detects the opposite of one object and another object based on one or more object identifiers that identify the opposing object. The opposite detection unit 16022 can be realized typically as an MPU, a memory, or the like. The processing procedure of the opposite detection means is usually realized by software, and the software is recorded on a recording medium such as a ROM. However, it may be realized by hardware (dedicated circuit).

  The link strength information acquisition unit 16023 acquires link strength information that is information indicating the frequency of speech detected by the speech detection unit 16021. Further, the link strength information acquisition unit 16023 acquires link strength information, which is information indicating the frequency of facing detected by the facing detection unit 16022. The link strength information acquisition unit 16023 may acquire both link strength information, which is information indicating the frequency of utterances, and link strength information, which is information indicating the frequency of facing, or one of them. The link strength information acquisition unit 16023 can usually be realized by an MPU, a memory, or the like. The processing procedure of the link strength information acquisition unit 16023 is usually realized by software, and the software is recorded in a recording medium such as a ROM. However, it may be realized by hardware (dedicated circuit).

  When the partial graph determination unit 105 determines that the partial graph satisfies a predetermined condition when the partial graph satisfies a predetermined condition, the summary extraction unit 16081 extracts the utterance portion constituting the partial graph from the content. In addition, when the partial graph determination unit 105 determines that the partial graph satisfies a predetermined condition, the summary extraction unit 16081 extracts, from the content, the opposite location constituting the partial graph. The summary extracting means 16081 can be usually realized by an MPU, a memory, or the like. The processing procedure of the summary extraction unit 16081 is usually realized by software, and the software is recorded in a recording medium such as a ROM. However, it may be realized by hardware (dedicated circuit).

The summary storage unit 16082 is a part of the content extracted by the summary extraction unit 16081, and stores summary content having audio or / and video of the part of the content. The storage medium to which the summary storage unit 16082 stores the summary content is preferably a nonvolatile recording medium such as a hard disk or a DVD-RAM. The summary storage unit 16082 may or may not include such a recording medium.
Next, the operation of the information processing apparatus will be described. Since the operation of the information processing apparatus is substantially the same as the operation of the information processing apparatus described in the first embodiment, the detailed operation description using the flowchart is omitted.
Hereinafter, a specific operation in which the information processing apparatus according to the present embodiment creates a video summary will be described.

  Now, assume that user A wears the information acquisition device, acquires the content shown in FIG. 18A, and transmits it to the information processing device. Further, it is assumed that the user B wears the information acquisition device, acquires the content shown in FIG. 18B, and transmits it to the information processing device.

  This content includes an “object identifier” for identifying the user, a video image taken by the CCD camera of the information acquisition device worn by the user, a user voice obtained by the microphone of the information acquisition device worn by the user, It has an object identifier of an object (here, a person or an exhibit) facing the user, acquired by the IR tracker of the information acquisition device that is attached.

Next, the information processing apparatus analyzes the two contents shown in FIGS. Here, it is assumed that the window width is “100 seconds”. Then, it is assumed that the information processing apparatus recognizes that it is an interaction that it is speaking to another person. In addition, the information processing apparatus configures a partial graph when the interaction is “10 seconds” or more in the window width of 100 seconds. In other words, when the link strength information is “10 (seconds)” or more, it is assumed that a new subgraph is recognized.
In the content of FIG. 18, the information processing apparatus does not detect an interaction (in this case, an utterance) when the time (t) is “0” to “100”.

  Next, the information processing apparatus detects an utterance from “A” to “B” between the times “100” and “200” of the content in FIG. 18A, and the link strength information acquisition unit 1602 Information on link strength information “80” is acquired. Also, the information processing apparatus detects an utterance from “B” to “A” between the times “160” and “260” of the content in FIG. 18B, and the link strength information acquisition unit 1602 Information of intensity information “90” is acquired.

  Then, the partial graph determination unit 105 determines that the link strength information “80” and “90” satisfy the condition “the link strength information is“ 10 (seconds) or more ””. That is, the subgraph determining unit 105 determines that the nodes “A” and “B” have shifted from the state of FIG. 19A (no interaction) to the state of FIG. 19B.

  Next, the partial graph information acquisition unit 107 configures a record of “ID = 1” in the partial graph information management table shown in FIG. Here, the partial graph information management table stores one or more pieces of partial graph information acquired or configured by the partial graph information acquisition unit 107. The partial graph information includes “object identifier”, “opposite object identifier”, “link strength information”, “start point (t)”, and “end point (t)”. Further, “ID” in the partial graph information management table is information for identifying a record and exists for record management in the table. The “object identifier” is information that identifies an object (here, a person) that is the starting point of an interaction, and the “opposite object identifier” is information that identifies an object (here, a person) that is subject to the interaction. “Link strength information” indicates the time of speech between two people. Further, “start point (t)” and “end point (t)” indicate the start time and end time of the utterance. From the record of “ID = 1” in the partial graph information management table, it can be seen that the users “A” and “B” are talking from the start point “100” to the end point “250”.

Next, the summary extraction means 16081 of the subgraph information processing unit 1608 has content (at least video and audio) in the dotted line frame of 181 in FIG. 18A based on the record of “ID = 1” in the subgraph information management table. ), The content of the dotted frame 182 in FIG. 18B (having at least video and audio) is acquired. Next, the summary storage unit 16082 stores the contents 181 and 182 of FIG. By such processing, content meaningful to the users “A” and “B” can be automatically stored only when the user “A” and the user “B” have a conversation.
Next, it is assumed that the contents of three users “A”, “B”, and “C” wearing the information acquisition apparatus are transmitted to the information processing apparatus and stored in the content storage unit 1601 of the information processing apparatus.

FIG. 21 is a diagram conceptually showing such content. FIG. 21A shows the content acquired by the information acquisition device of the user “A”. FIG. 21B shows the content acquired by the information acquisition device of the user “B”. FIG. 21C shows the content acquired by the information acquisition device of the user “C”. FIG. 21 shows contents after three times of time “1200” seconds.
Then, it is assumed that the information processing apparatus analyzes the content in FIG. 21 by the above-described operation and constructs a partial graph.

  That is, the information processing apparatus detects an utterance between the user “A” and the user “B” from the content 211 in FIG. 21A and the content 212 in FIG. Then, from the content 211 in FIG. 21A, the link strength information acquisition unit 1602 detects link strength information “80” from the node “A” to the node “B”. That is, the link strength information acquisition unit 1602 detects that the user “A” has spoken to the user “B” for 80 seconds. Also, from the content 212 in FIG. 21B, the link strength information acquisition unit 1602 detects link strength information “62” from the node “B” to the node “A”. That is, the link strength information acquisition unit 1602 detects that the user “B” has spoken to the user “A” for 62 seconds. Also, since “50” and “62” satisfy the above condition for constructing a subgraph, “10 (seconds)” or more, the subgraph information acquisition unit 107 displays the subgraph shown in FIG. get. Specifically, the subgraph information acquisition unit 107 configures and records the record “record ID = 1” in FIG. FIG. 23 is a partial graph information management table. The record of “record ID = 1” in FIG. 23 is information indicating a partial graph formed during the section (d) in FIG.

  Next, the information processing apparatus detects an utterance from the user “C” to the user “B”. That is, from the content 213 in FIG. 21C, the link strength information acquisition unit 1602 detects link strength information “38” from the node “C” to the node “B”. Also, from the content 214 of FIG. 21B, the link strength information acquisition unit 1602 detects link strength information “80” from the node “B” to the node “C”. Also, from the content 215 in FIG. 21C, the link strength information acquisition unit 1602 detects link strength information “47” from the node “C” to the node “A”. Then, the link strength information acquisition unit 1602 acquires all the link strength information of the link having the node “C” as the start point or the end point. Next, the subgraph connection strength information acquisition unit 103 acquires the subgraph connection strength information of the subgraphs of the nodes “A” and “B”. Specifically, the subgraph connection strength information acquisition unit 103 calculates the average of the link strength information “50” from the node “A” to “B” and the link strength information “62” from the node “B” to “A”. The value “56” is calculated as the subgraph connection strength information.

  Next, the partial graph determination unit 105 acquires the average value of the link strength information of the link whose node “C” is the start point or the end point. That is, the average value “61.25” of “80” “38” “80” “47” is calculated. Since “61.25” is larger than the subgraph connection strength information “52”, the subgraph determining unit 105 determines that the condition for adding the node “C” to the subgraph is met. To do. Here, the condition for adding one node to the subgraph is that the average value of the link strength information related to the one node is larger than the subgraph connection strength information.

  Next, the subgraph information acquisition unit 107 configures the record of “ID = 2” in FIG. 23 and writes it in the subgraph information management table. The record with “record ID = 2” in FIG. 23 is information indicating a partial graph formed during the section (e) in FIG. 21.

Then, the summary extraction unit 16081 of the subgraph information processing unit 1608 acquires the contents 213, 214, 215, and 216 of FIG. 21, and automatically extracts the summary contents including the conversations and videos of the three people. The summary storage unit 16082 stores the summary content.
Next, the contents of the four users “A”, “B”, “C”, and “D” wearing the information acquisition device are transmitted to the information processing device and stored in the content storage unit 1601 of the information processing device. To do.

FIG. 24 is a diagram conceptually showing such content. 24A, 24B, 24C, and 24D are contents acquired by the information acquisition apparatuses of the users “A”, “B”, “C”, and “D”, respectively. FIG. 24 shows the contents of the four contents after the time “3200” seconds.
Then, it is assumed that the information processing apparatus analyzes the content of FIG. 24 by the above-described operation and constructs a subgraph.

  That is, the information processing apparatus detects an utterance between the user “A” and the user “B” from the content 241 in FIG. 24A and the content 242 in FIG. Then, from the content 241 in FIG. 24A, the link strength information acquisition unit 1602 detects link strength information “50” from the node “A” to the node “B”. That is, the link strength information acquisition unit 1602 detects that the user “A” has spoken to the user “B” for 50 seconds. Also, from the content 242 in FIG. 24B, the link strength information acquisition unit 1602 detects link strength information “60” from the node “B” to the node “A”. That is, the link strength information acquisition unit 1602 detects that the user “B” has spoken to the user “A” for 60 seconds. Further, since “50” and “60” satisfy the above condition for constructing a subgraph, “10 (seconds)” or more, the subgraph information acquisition unit 107 displays the subgraph shown in FIG. get. Specifically, the partial graph information acquisition unit 107 configures and records a record indicating the above information in the partial graph information management table. In addition, since the example of partial graph information was mentioned above, detailed description here is abbreviate | omitted. Further, the information processing apparatus detects an utterance between the user “C” and the user “D” from the contents 243 and 245 in FIG. 24C and the content 244 in FIG. Then, from the contents 243 and 245 in FIG. 24C, the link strength information acquisition unit 1602 links the link strength information “50” from the node “C” to the node “D” (total time of the contents of 243 and 245). Is detected. Also, from the content 244 in FIG. 24D, the link strength information acquisition unit 1602 detects link strength information “40” from the node “D” to the node “C”. Also, since “50” and “40” satisfy the above condition for constructing a subgraph, “10 (seconds)” or more, the subgraph information acquisition unit 107 displays the subgraph shown in FIG. get. Then, the partial graph information acquisition unit 107 configures and records a record indicating the above information in the partial graph information management table.

  Next, the information processing apparatus detects an utterance from the user “C” to the user “B” from the content 246 in FIG. The link strength information acquisition unit 1602 detects link strength information “65” from the node “C” to the node “B”.

  Next, the subgraph connection strength information acquisition unit 103 acquires the subgraph connection strength information of the subgraphs of the nodes “A” and “B”. Specifically, the subgraph connection strength information acquisition unit 103 calculates the average of the link strength information “50” from the node “A” to “B” and the link strength information “60” from the node “B” to “A”. The value “55” is calculated as the subgraph connection strength information. Further, the subgraph connection strength information acquisition unit 103 acquires the subgraph connection strength information of the subgraphs of the nodes “C” and “D”. Specifically, the subgraph connection strength information acquisition unit 103 calculates the average of the link strength information “50” from the node “C” to “D” and the link strength information “40” from the node “D” to “C”. The value “45” is calculated as the subgraph connection strength information.

  Then, the subgraph determination unit 105 displays the subgraph connection strength information “55” of the subgraph in which the link strength information “65” from the node “C” to the node “B” is configured by the nodes “B” and “A”. It is detected that it is larger than the subgraph connection strength information “45” of the subgraph constituted by the nodes “C” and “D”. Then, the subgraph reconstruction control unit 106 instructs the subgraph determination unit 105 and the like to acquire a new subgraph among the nodes “A”, “B”, “C”, and “D”. Then, it is assumed that the partial graph information acquisition unit 107 has acquired partial graph information corresponding to a partial graph having “A”, “B”, and “C”, for example. An example of such partial graph information is shown in FIG. That is, FIG. 26 is an example of a partial graph information management table that stores the partial graph information that is reconfigured and corresponds to the newly detected partial graph. The record of “ID = 38” in FIG. 26 is subgraph information newly detected by reconstruction.

Next, the summary extraction unit 16081 of the subgraph information processing unit 1608 acquires the summary content corresponding to the record of “ID = 38” in FIG. 26 from the content storage unit 1601, and the summary storage unit 16082 stores the content. accumulate. With such processing, it is possible to automatically accumulate coherent content after a change has occurred in the subgraph.
Next, it is assumed that the contents of three users “A”, “B”, and “C” wearing the information acquisition apparatus are transmitted to the information processing apparatus and stored in the content storage unit 1601 of the information processing apparatus.

FIG. 27 is a diagram conceptually showing such content. 27A, 27B, and 27C are contents acquired by the information acquisition apparatuses of the users “A”, “B”, and “C”, respectively. FIG. 27 shows the contents after the time “5600” seconds of the three contents.
Then, it is assumed that the information processing apparatus analyzes the content in FIG. 27 by the above-described operation and constructs a partial graph.

  In other words, the information processing apparatus determines the user “A” from the contents 271, 272, and 273 in FIG. 27A, the contents 274 and 275 in FIG. 27B, and the contents 276 and 277 in FIG. The speech between “B” and “C” is detected. That is, the information processing apparatus extracts the partial graph shown in FIG. Since this process is the same as the process described above, detailed description thereof is omitted.

Next, the combined contribution information acquisition unit 104 determines each node from 5700 seconds to 5800 seconds (with a window width of 100) in the subgraph composed of the nodes “A”, “B”, and “C”. Acquires the contribution degree information of. That is, the combined contribution information acquisition unit 104 calculates the combined contribution information of the node “A” as “120” by “50 + 70” from the content 279 of FIG. Similarly, the connection contribution information of the node “B” is calculated as 120 ”. This “50” is the utterance time (second) from the user “A” to the user “B”, and “70” is the utterance time (second) from the user “B” to the user “A”.
Next, the combined contribution information acquisition unit 104 acquires combined contribution information “0” of the node “C” between 5700 seconds and after 5800 seconds.
Then, the subgraph connection strength information acquisition unit 103 acquires the subgraph connection strength information “120/3 = 40” of the subgraph composed of three nodes.

  Next, the subgraph determining unit 105 compares the subgraph connection strength information “40” with the connection contribution information “0”, and the connection contribution information “0” is determined to be ½ times the subgraph connection strength information. Since it is small, it is determined that the node “C” is deleted from the subgraph.

  Next, the subgraph information acquisition unit 107 acquires subgraph information that is information related to the subgraph excluding the node “C” from the subgraph having the nodes “A”, “B”, and “C”. Such partial graph information is a record of “ID = 80” in FIG. Then, the subgraph information acquisition unit 107 records such a record in the subgraph information management table.

Next, the summary extraction unit 16081 of the subgraph information processing unit 1608 extracts the content (contents 278 and 279 in FIG. 27) corresponding to the record of “ID = 80” in FIG. The summary storage unit 16082 stores the content.
As described above, according to the present embodiment, as a specific example of the processing described in the first embodiment, it is possible to automatically obtain a very effective summary video by using the video summary.
According to the present embodiment, the interaction is mainly utterance. However, opposing or other interactions are also possible.
Further, according to the present embodiment, the process of the subgraph information processing unit 108 is the creation of a video summary, but as described above, other processes may be used.

  Furthermore, the software that implements the information processing apparatus according to the present embodiment is the following program. That is, the program stores, in the computer, link strength information acquisition step for acquiring node information that is information about the node and link strength information that is information indicating the frequency of interaction between the nodes per predetermined time, and the link strength. Subgraph determination that determines whether or not a subgraph composed of two or more nodes and links between the nodes satisfies a condition related to a predetermined degree of connection based on one or more link strength information acquired in the information acquisition step A subgraph information acquisition step of acquiring subgraph information that is information related to the subgraph when the subgraph is determined to satisfy the predetermined condition in the step and the subgraph determination step; and the subgraph information A program for executing a subgraph information processing step that performs processing based on That.

  Further, in the above program, the node information is an object identifier for identifying an object, the interaction is a conversation between objects, and in the link strength information acquisition step, based on a voice included in one or more contents, An utterance detection step of detecting an utterance to another object from an object identified by one object identifier, and a link strength information acquisition step of acquiring link strength information which is information indicating the frequency of the utterance detected in the utterance detection step. And the subgraph processing step extracts, when the partial graph satisfies the predetermined condition, the portion of the utterance constituting the subgraph from the content when the subgraph is determined in the subgraph determination step. An extraction step and a summary extraction step A part of the out content, a program, comprising a summary storage step of storing the summary content with audio portion of the content.

  Further, in the above program, the node information is an object identifier for identifying an object, the interaction is an opposite between objects, and the link strength information acquisition step is an object identifier that two or more contents have, Based on one or more object identifiers that identify objects facing each other, a link detection step for detecting the facing of one object and another object, and information indicating the frequency of facing detected in the facing detection step A link strength information acquisition step for acquiring strength information, and when the subgraph satisfies the predetermined condition in the subgraph processing step, the subgraph is configured when determined in the subgraph determination step The opposite location from the content A summary extraction step of leaving a portion of the content extracted by the summary extraction step, a program, comprising a summary storage step of storing the summary content having a video portion of the content.

  FIG. 30 shows the external appearance of a computer that executes the programs described in this specification to realize the information processing apparatuses according to the various embodiments described above. The above-described embodiments can be realized by computer hardware and a computer program executed thereon. FIG. 30 is an overview diagram of the computer system 300, and FIG. 31 is a block diagram of the system 300.

  In FIG. 30, a computer system 300 includes a computer 301 including an FD (Flexible Disk) drive and a CD-ROM (Compact Disk Read Only Memory) drive, a keyboard 302, a mouse 303, a monitor 304, and a speaker 305. .

  In FIG. 31, in addition to the FD drive 3011 and the CD-ROM drive 3012, the computer 301 includes a CPU (Central Processing Unit) 3013, a bus 3014 connected to the CPU 3013, the CD-ROM drive 3012 and the FD drive 3011, and a boot. A ROM (Read-Only Memory) 3015 for storing a program such as an up program, and a RAM (Random Access Memory) connected to the CPU 3013 for temporarily storing instructions of the application program and providing a temporary storage space 3016 and a hard disk 3017 for storing application programs, system programs, and data. Although not shown here, the computer 301 may further include a network card that provides connection to a LAN.

  A program that causes the computer system 300 to execute the functions of the information processing apparatus according to the above-described embodiment is stored in the CD-ROM 3101 or FD 3102, inserted into the CD-ROM drive 3012 or FD drive 3011, and further stored in the hard disk 2517. May be forwarded. Alternatively, the program may be transmitted to the computer 301 via a network (not shown) and stored in the hard disk 3017. The program is loaded into the RAM 3016 at the time of execution. The program may be loaded directly from the CD-ROM 3101, the FD 3102, or the network.

  The program does not necessarily include an operating system (OS) or a third-party program that causes the computer 301 to execute the functions of the information processing apparatus according to the above-described embodiment. The program only needs to include an instruction portion that calls an appropriate function (module) in a controlled manner and obtains a desired result. How the computer system 300 operates is well known and will not be described in detail.

In each of the above embodiments, each process (each function) may be realized by centralized processing by a single device (system), or by distributed processing by a plurality of devices. May be.
In addition, the computer which performs the said program may be single, and plural may be sufficient as it. That is, centralized processing may be performed, or distributed processing may be performed.
The present invention is not limited to the above-described embodiments, and various modifications are possible, and it goes without saying that these are also included in the scope of the present invention.

  As described above, the information processing apparatus according to the present invention has an effect of being able to detect temporal changes in data having a graph structure composed of nodes and links, and is useful as an apparatus for automatically creating a video summary, for example. It is.

Block diagram of information processing apparatus according to Embodiment 1 Flow chart for explaining the operation of the information processing apparatus A flowchart for explaining the operation of the subgraph acquisition process A flowchart for explaining the operation of the change detection process A flowchart for explaining the operation of the subgraph generation determination process A flowchart for explaining the operation of the node addition determination process Flowchart explaining operation of interference determination process between subgraphs A flowchart for explaining the operation of the node deletion determination process A flowchart for explaining the operation of the update process of the subgraph information The figure explaining the processing concept of the information processing apparatus The figure explaining the processing concept of the information processing apparatus Figure showing the same partial graph information management table The figure explaining the processing concept of the information processing apparatus The figure explaining the processing concept of the information processing apparatus The figure explaining the processing concept of the information processing apparatus Block diagram of an information processing apparatus according to Embodiment 2 Overview of the information acquisition device The figure which shows the contents which are stored The figure explaining the processing concept of the information processing apparatus Figure showing the same partial graph information management table The figure which shows the contents which are stored The figure explaining the processing concept of the information processing apparatus Figure showing the same partial graph information management table The figure which shows the contents which are stored The figure explaining the processing concept of the information processing apparatus Figure showing the same partial graph information management table The figure which shows the contents which are stored The figure explaining the processing concept of the information processing apparatus Figure showing the same partial graph information management table The figure which shows the external appearance of the computer which implement | achieves the information processing apparatus Block diagram of a system constituting the information processing apparatus

Explanation of symbols

DESCRIPTION OF SYMBOLS 101 Node information storage part 102,1602 Link strength information acquisition part 103 Subgraph joint strength information acquisition part 104 Connection contribution information acquisition part 105 Subgraph judgment part 106 Subgraph reconstruction control part 107 Subgraph information acquisition part 108 Subgraph information Processing unit 1021 Interaction acquisition unit 1022, 16023 Link strength information acquisition unit 1601 Content storage unit 16021 Utterance detection unit 16022 Opposite detection unit 16081 Summary extraction unit 16082 Summary storage unit

Claims (8)

A link strength information acquisition unit that acquires node information that is information about the node, and link strength information that is information indicating the frequency of interaction between the nodes per predetermined time;
Based on one or more link strength information acquired by the link strength information acquisition unit, it is determined whether or not a subgraph composed of two or more nodes and a link between the nodes satisfies a condition regarding a predetermined degree of connectivity. A subgraph determination unit;
When the partial graph determination unit determines that the partial graph satisfies the predetermined condition, a partial graph information acquisition unit that acquires partial graph information that is information related to the partial graph;
An information processing apparatus comprising a partial graph information processing unit that performs processing based on the partial graph information. Based on the one or more link strength information acquired by the link strength information acquisition unit, it further comprises a subgraph connection strength information acquisition unit for calculating the degree of coupling of the subgraphs and acquiring the partial graph connection strength information,
The partial graph determination unit
The information processing apparatus according to claim 1, wherein the information processing apparatus determines whether or not the partial graph satisfies a condition relating to a predetermined degree of connection based on the link strength information and the partial graph connection strength information. Based on link strength information between one node of two or more nodes constituting the subgraph and another node, the connection contribution is information indicating the contribution degree of the connection in the subgraph of the one node Further comprising a combined contribution information acquisition unit for calculating a degree and acquiring combined contribution information,
The partial graph determination unit
Based on the link strength information and the connection contribution information, or the link strength information, the subgraph connection strength information, and the connection contribution information, it is determined whether or not the partial graph satisfies a condition regarding a predetermined connection degree. The information processing apparatus according to claim 1. The link strength information acquisition unit
Obtaining link strength information between nodes corresponding to node information that does not exist in the node information included in the one or more subgraph information stored;
The partial graph determination unit
The information processing apparatus according to claim 1, wherein the link strength information acquired by the link strength information acquisition unit determines whether the link strength information is information indicating a degree of coupling greater than or equal to a predetermined value. The link strength information acquisition unit
The first node corresponding to one node information in the node information included in the one or more stored partial graph information and the node information included in the one or more stored partial graph information do not exist Obtain link strength information between the second node corresponding to the node information,
The partial graph determination unit
When the link strength information acquired by the link strength information acquisition unit and the partial graph connection strength information of the subgraph having the first node satisfy a predetermined condition, the second node is added to the subgraph. The information processing apparatus according to claim 2, wherein the subgraph is determined to satisfy a condition regarding a predetermined degree of coupling. The link strength information acquisition unit
The first node corresponding to one node information in the node information included in one stored partial graph information, and one node information in the node information included in the other stored partial graph information To obtain link strength information between the second nodes corresponding to
The partial graph determination unit
The link strength information acquired by the link strength information acquisition unit and the partial graph connection strength information of the first partial graph corresponding to the one partial graph information satisfy a predetermined condition, and the link strength acquired by the link strength information acquisition unit Determining whether the subgraph connection strength information of the second subgraph corresponding to the information and the other subgraph information satisfies a predetermined condition;
When the partial graph determination unit determines that the predetermined condition is satisfied, the link strength information acquisition unit is instructed to acquire link strength information between nodes constituting the first partial graph and the second partial graph. Or instructing the link strength information acquisition unit to acquire link strength information between nodes constituting the first subgraph and the second subgraph, and subgraph connection strength information acquisition unit The information processing according to claim 2, further comprising: a subgraph reconstruction control unit that instructs to acquire information and subsequently instructs the subgraph determination unit to determine whether or not a predetermined condition is satisfied. apparatus. The partial graph determination unit
Whether or not the contribution contribution degree information of the first node corresponding to one node information in the node information included in one stored subgraph information and the subgraph connection strength information of the subgraph satisfy a predetermined condition Determine whether
The partial graph information acquisition unit
The information processing according to claim 3, wherein when the subgraph determining unit determines that the predetermined condition is satisfied, the subgraph information is acquired from the subgraph, which is information related to the subgraph excluding the first node. apparatus. On the computer,
A link strength information acquisition step for acquiring node information that is information about the node, and link strength information that is information indicating the frequency of interaction between the nodes per predetermined time;
Based on one or more link strength information acquired in the link strength information acquisition step, it is determined whether or not a subgraph composed of two or more nodes and a link between the nodes satisfies a predetermined connection degree condition. A subgraph determination step;
In the subgraph determination step, when it is determined that the partial graph satisfies the predetermined condition, a subgraph information acquisition step of acquiring subgraph information that is information related to the subgraph;
A program for executing a subgraph information processing step for performing processing based on the subgraph information.

Images