JP2006251042A - Information processor, information processing method and program - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To provide an information processor by which the circumstances of a conference can simply be known. <P>SOLUTION: A conference circumstances analysis apparatus 1 is provided with an uttering database 2 which stores conference data including uttering data, an analysis section 3 which adds up the number of uttered words from the uttering data stored in the uttering database 2 to obtain an average of the number of the uttered words and a standard deviation of the number of the uttered words, an annotation database 4 which stores the average of the number of the uttered words and the standard deviation of the uttered words obtained by the analysis section 3 while relating them to the uttering data and a display control section 5 which displays a graph in which the average of the number of the uttered words and the standard deviation of the number of the uttered words are made to correspond to the time axis of the uttering data. Thus, the respective amount of uttering and its leaning of the conference attendees can be computed and a large leaning case in which a single person utters continuously and a small leaning case in which a plurality of persons are uniformly uttering can be determined. <P>COPYRIGHT: (C)2006,JPO&NCIPI

Description

  The present invention relates to an information processing apparatus, an information processing method, and a program.

  Conventionally, in a conference, it is convenient to be able to display the mutual relations of the comments that are related to each other on the speaker chart. For example, when one conference participant gives an opinion to another conference participant or asks a question, an interactive scene in which another conference participant responds or refutes is shown on the chart. If it can be discriminated, it will be understood that there was some discussion, and it is expected that the searcher of the recorded conference information can easily search for the discussion part to be reproduced by using it as a clue. The following are proposed as conventional techniques for realizing this.

  The device described in Patent Document 1 records information on a speech section for each speaker and an attitude for each speaker in association with each other, further identifies who the speech is for, and a speech in a predetermined time section. Since the flow of the message is displayed, the relationship between the statements can be displayed.

JP-A-11-259501

  However, although the above-mentioned prior art can know the relationship between the statements, a presentation scene where a single speaker continues to speak for a while or a discussion scene where many participants speak uniformly. As such, I could not know the mode of the conference, such as how the speakers are biased and excited. Therefore, the meeting record data had to be confirmed while flowing from the beginning to the end.

  Therefore, the present invention has been made in view of the above problems, and an object thereof is to provide an information processing apparatus, an information processing method, and a program capable of easily knowing the state of a conference.

  In order to solve the above problems, the present invention provides a storage means for storing utterance data, the number of utterance words from the utterance data stored in the storage means, an average of the utterance words, and a standard of the number of utterance words It is an information processing apparatus provided with the analysis means which calculates | requires at least one of a deviation. According to the present invention, the number of utterance words is totaled from the utterance data, and at least one of the average number of utterance words and the standard deviation of the number of utterance words is obtained. Since it is possible to calculate, it is possible to determine a case where there is a large bias such as when one person is uttering all the time or a case where there is a small bias such as when a plurality of people are speaking equally. In this way, for example, a presentation scene where a single speaker has been speaking for a while, a discussion scene where many participants are speaking uniformly, etc. You can easily know the mode of the meeting.

  The information processing apparatus according to the present invention further includes storage means for storing at least one of the average number of utterance words obtained by the analysis means and the standard deviation of the number of utterance words in association with the utterance data. The information processing apparatus according to the present invention includes a storage unit that stores an average of the number of utterance words obtained by the analysis unit or a standard deviation of the number of utterance words in association with the utterance data, and a search input from the storage unit Search means for searching for at least one of the average number of spoken words satisfying the condition and the standard deviation of the number of spoken words is further provided.

  The information processing apparatus of the present invention further includes display control means for displaying a graph in which at least one of the average number of spoken words and the standard deviation of the number of spoken words is associated with the time axis of the speech data. The analysis means assigns at least one of the average of the number of spoken words and the standard deviation of the number of spoken words as an annotation of the speech data. Thereby, a search can be performed later based on the assigned annotation.

  The present invention includes a step of identifying a speaker based on utterance data cut out at a predetermined time, and a step of counting the number of utterance words of each speaker based on the utterance data cut out at the predetermined time; And obtaining at least one of an average of the number of spoken words and a standard deviation of the number of spoken words based on the number of spoken words of each speaker. According to the present invention, the number of utterance words is totaled from the utterance data, and at least one of the average number of utterance words and the standard deviation of the number of utterance words is obtained. Since it is possible to calculate, it is possible to determine a case where there is a large bias such as when one person is uttering all the time or a case where there is a small bias such as when a plurality of people are speaking equally.

  The information processing method of the present invention further includes a step of displaying a graph in which at least one of the average number of spoken words and the standard deviation of the number of spoken words is associated with the time axis of the utterance data. The information processing method of the present invention further includes a step of assigning at least one of an average of the number of speaker words and a standard deviation of the number of speaker words as an annotation of the utterance data.

  The present invention includes a step of identifying a speaker based on utterance data cut out at a predetermined time, a step of counting the number of utterance words of each speaker based on the utterance data cut out at the predetermined time, A program for causing a computer to execute a step of obtaining at least one of an average of spoken words and a standard deviation of spoken words based on the number of spoken words of a speaker. According to the present invention, the number of utterance words is totaled from the utterance data, and at least one of the average number of utterance words and the standard deviation of the number of utterance words is obtained. Since it is possible to calculate, it is possible to determine a case where there is a large bias such as when one person is uttering all the time or a case where there is a small bias such as when a plurality of people are speaking equally.

  The program according to the present invention further causes the computer to generate information for displaying a graph in which at least one of the average number of spoken words and the standard deviation of the number of spoken words is associated with the time axis of the speech data Let

  According to the present invention, it is possible to provide, for example, an information processing apparatus, an information processing method, and a program that can easily know the mode of a meeting.

  Hereinafter, the best mode for carrying out the present invention will be described.

  FIG. 1 is a block diagram of a conference mode analysis apparatus (information processing apparatus) according to a first embodiment of the present invention. As shown in FIG. 1, the meeting state analysis apparatus 1 includes an utterance database 2, an analysis unit 3, an annotation database 4, a display control unit 5, and a display unit 6. The utterance database 2 stores conference record data including at least voice data (utterance data) of conference participants.

  FIG. 2 is a diagram showing the contents of the utterance database 2. As shown in FIG. 2, the utterance database 2 includes fields of time, speaker (A, B, C, D), and utterance content. In FIG. 2, the content of the utterance is shown in text, but what is actually stored in the utterance database 2 is voice data. The analysis unit 3 totals the number of utterance words of each speaker from the utterance data stored in the utterance database 2 to obtain at least one of the average value of the utterance words and the standard deviation of the number of utterance words. The average and the standard deviation of the number of spoken words are given as annotations of the speech data. Here, the standard deviation of the number of spoken words is an average of the distance (distance) from the average of the number of spoken words.

  The annotation database 4 stores the average of the number of utterance words and the standard deviation of the number of utterance words obtained by the analysis unit 3 in association with the utterance data using the time. The display control unit 5 controls the display, and displays a graph in which the average number of utterance words of each speaker and the standard deviation of the number of utterance words are associated with the time axis of the utterance data. The display unit 6 is configured by a display device, for example.

  Next, the operation of the analysis unit 3 will be described. FIG. 3 is an operation flowchart of the analysis unit 3. In step S11, the analysis unit 3 cuts out data for a predetermined time (for example, 1 minute) from the utterance database 2. In step S12, the analysis unit 3 identifies speakers appearing like speaker B, speaker C, and speaker E, for example. In step S <b> 13, the analysis unit 3 counts the number of utterance words per minute for each speaker. Here, for example, it is assumed that the number of utterance words of speaker B is 50 words, the number of utterance words of speaker C is 5 words, and the number of utterance words of speaker E is 5 words. In step S14, the analysis unit 3 calculates the average value X of the number of spoken words. In addition, the analysis unit 3 calculates the standard deviation σ of the number of utterance words using Expression 1.

ここで、Ｎは話者数、ｘは各話者の発話単語数を示す。ステップＳ１５で、発話単語数の平均値Ｘと発話単語数の標準偏差σをアノテーションデータベース４に保存する。

Here, N represents the number of speakers, and x represents the number of words spoken by each speaker. In step S15, the average value X of the number of spoken words and the standard deviation σ of the number of spoken words are stored in the annotation database 4.

  FIG. 4 is a display example by the display unit 6. In FIG. 4, the horizontal axis corresponds to the conference start time and conference end time of the conference record data stored in the utterance database 2 as the conference elapsed time. In the vertical axis, the broken line indicates the average value X of the number of spoken words of each speaker, and the solid line indicates the standard deviation σ of the number of spoken words. When the average value X of the number of spoken words is large and the standard deviation σ of the number of spoken words is small, it can be seen that the participants are speaking evenly (1). When the average value X of the number of spoken words is small and the standard deviation σ of the number of spoken words is small, it can be seen that everyone is not speaking (2). When the average value X of the number of uttered words is large and the standard deviation σ of the number of uttered words is large, it can be seen that one person is speaking in a concentrated manner (3). If one point on the time axis is selected, the sound there can be reproduced. Also, by moving or clicking the bar 20 indicated by the dotted line, the speaker data (voice) at that time can be reproduced. As described above, by analyzing the voice of the participant, for example, a presentation scene where a single speaker has been speaking for a while, a discussion scene where many participants are speaking uniformly, etc. It is possible to determine the state of the conference such as how the speakers are biased and excited.

  Next, a second embodiment will be described. FIG. 5 is a block diagram of a conference mode search apparatus (information processing apparatus) according to the second embodiment of the present invention. As shown in FIG. 5, the meeting state search device 100 includes an utterance database 2, an analysis unit 3, an annotation database 4, a display control unit 5, a display unit 6, a search condition setting unit 7, and a search unit 8. In addition, the same part as Example 1 attaches | subjects the same code | symbol, and abbreviate | omits description. The utterance database 2 stores conference record data including at least voice data (utterance data) of conference participants for a plurality of conferences.

  The analysis unit 3 aggregates the number of utterance words of each speaker from the utterance data stored in the utterance database 2 for each conference, and obtains at least one of the average value of the number of utterance words and the standard deviation of the number of utterance words. . The annotation database 4 stores the average of the number of utterance words and the standard deviation of the number of utterance words obtained by the analysis unit 3 for each conference (conference 1 to conference n) in association with the utterance data using time. The annotation database 4 stores one set of data in the format of time, average number of spoken words (average amount of spoken words), and standard deviation of the number of spoken words for one meeting record data.

  The search condition setting unit 7 is composed of input means such as a mouse and a keyboard, for example, “parts where the participants speak evenly”, “parts where no one speaks”, “one person speaks intensively” It is possible to set a meeting mode such as “a part that is present”, and it is also possible to directly specify the average of spoken words and the standard deviation of spoken words. Thereby, it is also possible to search the conference record data using the conference mode as a key. The search unit 8 searches the annotation database 4 for the average number of utterance words that satisfy the input search condition and the standard deviation of the number of utterance words. Here, it is assumed that conference 3 and conference 5 are hit.

  As shown in FIG. 6, the display control unit 5 displays a graph in which the average number of utterance words and the standard deviation of the number of utterance words in the conference 3 and the conference 5 are associated with the time axis of the utterance data on the display unit 6. Furthermore, the number of utterance words and the standard deviation of the number of utterance words of the conference 3 and the conference 5 are displayed in an identifiable manner.

  According to the conference mode retrieval apparatus according to the present embodiment, a set of data as described above is created and stored for a large amount of conference record data. It is possible to search for a desired scene using a conference mode such as a part where participants are speaking equally, a part where everyone is not speaking, or a part where one person is speaking intensively.

  The information processing method according to the present invention is realized using, for example, a CPU (Central Processing Unit), a ROM (Read Only Memory), a RAM (Random Access Memory), and the like, and the program is stored in a hard disk device, a CD-ROM, or a DVD. Alternatively, each step is realized by installing from a portable storage medium such as a flexible disk or downloading from a communication circuit and the CPU executing this program. That is, the program specifies a speaker based on utterance data cut out at a predetermined time, totals the number of utterance words of each speaker based on the utterance data cut out at the predetermined time, Determining at least one of an average of the number of spoken words and a standard deviation of the number of spoken words based on the number of spoken words of each speaker, and calculating at least one of the average of the number of spoken words and the standard deviation of the number of spoken words of the speech A step of generating information for displaying a graph corresponding to the time axis of data is executed by a CPU (computer).

  Although the preferred embodiments of the present invention have been described in detail above, the present invention is not limited to the specific embodiments, and various modifications, within the scope of the gist of the present invention described in the claims, It can be changed.

It is a block diagram of the meeting state analysis apparatus which concerns on 1st Example of this invention. It is a figure which shows the content of the speech database. 5 is an operation flowchart of the analysis unit 3. It is a display example by the display unit 6. It is a block diagram of the meeting mode search apparatus which concerns on 2nd Example of this invention. It is a figure which shows the graph which matched the average of the number of utterance words of the meeting 3 and the meeting 5, and the standard deviation of the number of utterance words with the time-axis of utterance data.

Explanation of symbols

DESCRIPTION OF SYMBOLS 1 Meeting state analysis apparatus 2 Utterance database 3 Analysis part 4 Annotation database 5 Display control part 6 Display part 100 Meeting state search apparatus 7 Search condition setting part 8 Search part

Claims (10)

Storage means for storing utterance data;
An information processing apparatus comprising: an analysis unit that aggregates the number of utterance words from the utterance data stored in the storage unit and obtains at least one of an average of the number of utterance words and a standard deviation of the number of utterance words. The information processing apparatus according to claim 1, further comprising a storage unit that stores at least one of an average of the number of utterance words obtained by the analysis unit and a standard deviation of the number of utterance words in association with the utterance data. . Storage means for storing an average of the number of utterance words obtained by the analysis means or a standard deviation of the number of utterance words in association with the utterance data;
The information according to claim 1, further comprising: search means for searching for at least one of an average of the number of spoken words satisfying a search condition input from the storage means and a standard deviation of the number of spoken words. Processing equipment. 4. The display control means for displaying a graph in which at least one of the average number of utterance words and the standard deviation of the number of utterance words is associated with the time axis of the utterance data. The information processing apparatus described in 1. The information processing apparatus according to claim 1, wherein the analysis unit assigns at least one of an average of the utterance words and a standard deviation of the utterance words as an annotation of the utterance data. Identifying a speaker based on utterance data cut out at a predetermined time;
Based on the utterance data cut out in the predetermined time, the step of counting the number of utterance words of each speaker;
Obtaining at least one of an average of the number of spoken words and a standard deviation of the number of spoken words based on the number of spoken words of each speaker. The information processing method according to claim 6, further comprising a step of displaying a graph in which at least one of the average number of spoken words and the standard deviation of the number of spoken words is associated with a time axis of the speech data. . The information processing method according to claim 6, further comprising a step of assigning at least one of an average of the number of speaker words and a standard deviation of the number of speaker words as an annotation of the utterance data. Identifying a speaker based on utterance data cut out at a predetermined time;
A step of counting the number of utterance words of each speaker based on the utterance data cut out at the predetermined time;
A program for causing a computer to execute a step of obtaining at least one of an average of the number of spoken words and a standard deviation of the number of spoken words based on the number of spoken words of each speaker. The computer further executes a step of generating information for displaying a graph in which at least one of the average number of spoken words and the standard deviation of the number of spoken words is associated with a time axis of the speech data. Program.

Images