JP5751143B2 - Minutes creation support device, minutes creation support system, and minutes creation program - Google Patents

Description

  The present invention relates to a minutes creation support apparatus, a minutes creation support system, and a minutes creation program, and more particularly, a minutes creation support apparatus, a minutes creation support system, and a minutes creation that can extract important statements. Related to the program.

Conventionally, various techniques have been disclosed for creating minutes.
For example, Patent Document 1 (Japanese Patent Laid-Open No. 2007-215053) discloses a technique for creating minutes during a meeting. According to the technology described in Patent Document 1, when the minutes are created, it is set as the start time of an important statement, and the important statement ends with the meeting end time.

  Non-Patent Document 1 ("VoiceGraphy-Product Overview", [online], NEC Corporation, [searched on September 21, 2011], Internet <http://www.nec.co.jp/soft/ In VoiceGraphy / outline.html>), the minutes are generated by text recognition by voice recognition.

  Also, Non-Patent Document 2 ("Discussion Mining Project", [online], Department of Media Science, Graduate School of Information Science, Nagoya University, [Search September 21, 2011], Internet <http: //www.nagao.nuie .nagoya-u.ac.jp / projects / discussion.xml>) contains real-world activities (for example, meetings) using multiple cameras, microphones, and browser-based tools. The presenter automatically records this information in the minutes by using a dedicated tool to convey the slides used for the presentation and the timing for switching the slides. In addition, the content of a statement is recorded by the clerk using a dedicated tool. More specifically, when a participant speaks, a tag-type device called a discussion tag is used to convey the speaker's ID and the statement type (question, answer, comment) to the secretary. Supports the creation of records and helps to structure the discussion. In such a system, information such as the slides used and URLs (Uniform Resource Locator) of Web materials are automatically captured in the minutes, and the discussion itself generates a structured minutes. It has become.

  In addition, as a technique for recording conversations between participants at an event or the like, Non-Patent Document 3 ("What is Business Microscope (registered trademark)", [online], Hitachi, Ltd., [September 21, 2011 Search], the Internet <http://www.hitachi-hitec.com/jyouhou/business-microscope/solution/microscope.html>), a technology called a business microscope is disclosed. In this technology, an ID card type sensing device is used. The apparatus includes a microphone and a face-to-face sensor. This microphone captures the voice and environmental sound of the user wearing the device. In addition, the face-to-face sensor identifies the faced person by communicating with an ID card type sensing device worn by the faced person.

JP 2007-215053 A

"VoiceGraphy-Product Overview", [online], NEC Corporation, [searched on September 21, 2011], Internet <http://www.nec.co.jp/soft/VoiceGraphy/outline.html> "Discussion Mining Project", [online], Department of Media Science, Graduate School of Information Science, Nagoya University, [Search September 21, 2011], Internet <http://www.nagao.nuie.nagoya-u.ac. jp / projects / discussion.xml> "What is a business microscope (registered trademark)", [online], Hitachi, Ltd., [searched on September 21, 2011], Internet <http://www.hitachi-hitec.com/jyouhou/business-microscope /solution/microscope.html>

  In recent years, the life log has a mechanism that saves and uses all sorts of miscellaneous information such as audio data that has been cut into frames, documents created on the spot (such as meetings), and whiteboard images. Proposed. Various studies have been made on the use of a system that uses such life logs to extract human behavior patterns and connect them to marketing.

  In creating the minutes of events such as conferences, it is useful to extract important parts for the event from the life log as described above. It takes time and effort to prepare the minutes manually from the recorded records when the minutes maker observes the entire event or only records the entire event. Because.

  However, according to the technique described in Patent Document 1, the timing of appearance of important utterances is set corresponding to the conference start time, but there are cases where important utterances cannot be recorded by such a method. Furthermore, it is assumed that minutes will be created during the meeting and cannot be created later. Therefore, it is not possible to prepare for the minutes after the meeting.

  Further, according to the technique described in Non-Patent Document 1, it is possible to obtain a list of statements in an event, but the point of extracting important parts is not taken into consideration.

  Further, according to the technique described in Non-Patent Document 2, it is possible to structure and analyze the discussion that occurred during the event, but it requires a speaker to operate a bill-type device, which is complicated for the participants. . Also, if the speaker forgets to operate the device and makes a statement, the minutes cannot be created. Furthermore, it is considered difficult to apply in a situation where participants are not particularly aware of the minutes, such as conversations in daily work.

  Further, according to the technique described in Non-Patent Document 3, although a correlation diagram between users is created, the point of extracting an important part of an event has not been studied.

  The present invention has been conceived in view of such circumstances, and its purpose is to extract a minutes of a meeting by extracting an important part of the meeting, a meeting creation support system, and It is to provide a program for creating minutes.

A minutes creation support apparatus according to an aspect of the present invention includes a detection unit for detecting a state in which a conversation between conference participants is established, and a record for recording a voice uttered by the conference participants. And a means for extracting a specific voice from the voice recorded in the recording means based on the detection result of the conversation establishment state by the means, and a meeting minutes using the specific voice And a means for creating a minutes for the detection, and the detection means is a set of participants who can communicate with each other of the infrared transmission / reception unit of the device having an individual ID attached by each participant and have received each other's individual ID. but if that is present, it judges that it is the state where the conversation among participants of the conference is established.
A minutes creation support apparatus according to another aspect of the present invention is for detecting a state in which a conversation between participants in a conference is established, and for recording a sound emitted by a participant in the conference. Recording means, extraction means for extracting specific sound from the sound recorded in the recording means based on the detection result of the conversation establishment state by the detecting means, and creating the meeting minutes using the specific sound And a means for creating a minutes for detecting the position and orientation of each participant by means of an electronic compass and a GPS controller of a device worn by each participant; and When voice is detected in a pair of participants facing each other, it is determined that a conversation between the participants in the conference is established.

  Preferably, for each participant in the conference, the extraction means calculates the total conversation time by adding up the times when the conversation is determined to be established by the detection means, and based on the total conversation time for each participant, The main speaker in the conference is specified, and the voice of the conversation including the main speaker is extracted as the specific voice from the voice recorded by the recording means.

  Preferably, the extraction means calculates the conversation concentration level of each participant by calculating the ratio of the total conversation time of each participant to the total conversation time for all participants, and the conversation concentration level of each participant. Based on the above, the participant having the highest conversation concentration is determined as the main speaker, and the voice of the conversation including the main speaker is extracted as the specific voice from the voice recorded by the recording means.

  Preferably, the recording unit records the sound by distinguishing between the sound uttered by each participant and the sound uttered by the participant.

  Preferably, the minutes creation unit generates a link for a specific voice and presents the link.

  Preferably, the extraction unit extracts a specific voice from the voice recorded in the recording unit based on the priority based on the time zone of the voice utterance.

  Preferably, the extraction unit extracts specific audio from the audio recorded in the recording unit based on the priority set for the participant who has emitted the audio recorded in the recording unit.

Preferably, the recording means records the sound together with the biological information of the person who emits the sound .

A minutes creation support system according to an aspect of the present invention is a minutes creation support system including a detection means worn by a conference participant and a minutes creation support device capable of communicating with the detection means. The minutes preparation support device, based on the detection result obtained from each detection means, a detection means for detecting a state in which the conversation between the conference participants is established, and the sound emitted by the conference participants A recording means for recording, an extraction means for extracting a specific sound from the sound recorded in the recording means based on the detection result of the conversation establishment state by the detecting means, and a meeting using the specific sound look including the proceedings creation means for creating the minutes, detection means becomes each communicable infrared transceiver devices with individual ID worn by each participant, the individual ID of the other party mutually Receive If the set of participants are present, it judges that it is the state where the conversation among participants of the conference is established.
A minutes creation support system according to another aspect of the present invention is a minutes creation support system comprising a detection means worn by a conference participant and a minutes creation support device capable of communicating with the detection means. Thus, the minutes creation support device, based on the detection results obtained from each detection means, a detection means for detecting a state in which a conversation between the conference participants is established, and a sound emitted by the conference participants A recording means for recording the voice, an extraction means for extracting a specific voice from the voice recorded in the recording means based on the detection result of the conversation establishment state by the detection means, and a meeting using the specific voice The minutes detection means for creating the minutes of the device, and the detection means identify the position and orientation of each participant by the electronic compass and GPS controller of the device worn by each participant, and in the vicinity Place And, and, when it is detected voice in the set of participants who are opposed, it is determined that the state in which the conversation between the participants of the conference is established.

A minutes creation program according to an aspect of the present invention is a computer-readable program, and a computer detects a state in which a conversation between participants in a conference is established. Recording means for recording the voice uttered by the participant, extraction means for extracting a specific voice from the voice recorded in the recording means based on the detection result of the conversation establishment state by the detection means, and a specific It functions as a minutes creation means for creating the minutes of the meeting using voice, and the detection means enables each of the infrared transmission / reception units of the devices having individual IDs attached by each participant to communicate with each other. When there is a set of participants who have received the other party's individual ID, it is determined that a conversation between the participants in the conference has been established .
A minutes creation program according to another aspect of the present invention is a computer-readable program, and a computer detects a state in which a conversation between conference participants is established, a meeting Recording means for recording the voice uttered by the participants, extraction means for extracting a specific voice from the voice recorded in the recording means based on the detection result of the conversation establishment state by the detecting means, and a specific It functions as a minutes creation means for creating the meeting minutes using the voice of the device, and the detection means detects the position and orientation of each participant by the electronic compass and GPS controller of the device worn by each participant. In particular, if audio is detected in a pair of participants who are located in the vicinity and face each other, it is determined that the conversation between the participants in the conference has been established. To.

  The conversation is recorded, the establishment state of the conversation is detected, and based on the result of detecting the establishment, a specific sound is extracted from the recorded sound to create a minutes.

  As a result, a conversation is established, and further, a voice to be used for creating the minutes is extracted according to how the conversation is established.

  As a result, it is possible to extract the audio considered important in the conference and create the minutes.

It is a figure which shows schematic structure of the minutes creation assistance system which is one embodiment of this invention. It is a figure which shows typically an example of a function structure of a minutes creation assistance system. It is a figure which shows an example of the external appearance of a conversation collection device. It is a figure which shows typically an example of the hardware constitutions of a conversation collection device. It is a figure which shows an example of the hardware constitutions of a server typically. It is a figure which shows an example of the scenery of a meeting typically. It is a figure which shows typically an example of the conversation state produced | generated in one embodiment of this invention. It is a figure which shows typically an example of the conversation time of each participant produced | generated in one embodiment of this invention. It is a figure for demonstrating the calculation method of a conversation concentration degree. It is a figure which shows typically the content of the audio | voice data corresponding to the utterance extracted by the extraction part among the utterances of each person shown by FIG. It is a flowchart of the minutes creation process performed in the minutes creation system of one embodiment of this invention. It is a figure which shows an example of the tag information registered beforehand for every meeting. It is a figure which shows an example of the screen displayed on the display device of a server. It is a figure which shows the example of a display of a minutes file. It is a figure which shows an example of the screen for displaying the information linked to each audio | voice data. It is a figure which shows the external appearance of the modification of a conversation collection device. It is a figure which shows typically an example of the hardware constitutions of the conversation collection device of FIG. It is a figure which shows typically an example of the hardware constitutions of the other modification of a conversation collection device. It is a figure which shows the modification of the display screen of the link of FIG. It is a figure which shows typically an example of the hardware constitutions of the other modification of a conversation collection device.

  The minutes creation support system according to the present invention will be described below with reference to the drawings. In addition, in each figure, the same code | symbol is attached | subjected about the component which show | plays the same effect | action and function, and the description is not repeated.

[Schematic configuration of the minutes creation system]
FIG. 1 is a diagram showing a schematic configuration of a minutes creation support system according to an embodiment of the present invention.

  Referring to FIG. 1, the minutes creation support system communicates with each of conversation collection devices 200A to 200D and conversation collection devices 200A to 200D, which are worn by each user (participant of the conference) and function as detection means. And a server 100 functioning as a minutes creation support apparatus.

  In this specification, the conversation collection devices 200 </ b> A to 200 </ b> D may be collectively referred to as “conversation collection device 200”. The number of conversation collection devices may be the same as the number of participants, or less. That is, all the participants in the conference may be equipped with the conversation collection device, or only some of the participants may be equipped with the conversation collection device.

  The conversation collection device is an apparatus for acquiring information for detecting a state in which a conversation is established between the participants in the conference, and acquiring voice (conversation) uttered by the participants. Therefore, in the minutes creation support system, if a device capable of acquiring such information about a plurality of participants is provided, it is not necessary to provide the same number as the participants even for acquiring information of all participants. In some cases, it may be singular.

FIG. 2 is a diagram schematically illustrating an example of a functional configuration of the minutes creation support system.
The minutes creation support system mainly includes a life log generation system 900 and a minutes creation system 800. The life log generation system 900 includes one or more conversation collection devices 200, a tag generation unit 910, and a data storage unit 920, and has a function of acquiring and storing data of participants in the conference during the conference. The tag generation unit 910 generates tag information about the conference.

  Participant data acquired via the conversation collection device 200 is stored in the data storage unit 920 together with the tag information generated in the tag generation unit 910. More specifically, the conversation collection device 200 acquires a part for acquiring the participant's own information such as the voice of each participant, and information for specifying who each participant is talking to. And having a part. These pieces of information are stored together with tag information attached to the conference so that the search can be performed after the conference ends. The tag generation unit 910 gives date information, conference information from a scheduler (not shown), and the like as tag information. In the life log generation system 900, the information generated in real time in this way is stored in the data storage unit 920.

  The minutes creation system 800 includes a search unit 810, an input unit 820, a specific speaker identification unit 830, a conversation concentration calculation unit 840, an extraction unit 850, a priority setting unit 860, and a link generation unit 870. And a minutes creation unit 880. In the present embodiment, the conversation concentration level calculation unit 840 functions as a detection unit, and the specific speaker identification unit 830, the conversation concentration level calculation unit 840, and the extraction unit 850 function as an extraction unit.

  The search unit 810 searches the data storage unit 920 for tag information of a meeting for which minutes are to be created. The input unit 820 accepts input of information such as a minutes creation instruction. Note that the input unit 820 may receive an instruction for creating the minutes from an external device via a network. Note that the input unit 820 may accept input of information for generating tag information.

  The conversation concentration degree calculation unit 840 calculates a conversation concentration degree described later for each participant in the conference. The specific speaker specifying unit 830 specifies a participant who uses the speech for creating the minutes based on the conversation concentration level and the like from the participants of the conference. The extraction unit 850 extracts, from the data storage unit 920, the conversation involving the participant specified by the specific speaker specifying unit 830. The priority setting unit 860 sets the priority for creating the minutes for the participants of the conference. The link generation unit 870 generates link information from the top page of the minutes file to various data such as voice data that constitutes the minutes. The minutes creation unit 880 creates a minutes file by associating the top page with the link information as described above.

[Conversation Collection Device Configuration]
FIG. 3 is a diagram illustrating an example of the appearance of the conversation collection device 200.

  Referring to FIG. 3, conversation collecting device 200 has an ID card type form that a conference participant wears from the neck. The conversation collection device 200 includes a microphone 201 and an infrared transmission / reception unit 202. The microphone 201 can acquire the sound emitted by the wearer.

  The conversation collection device 200 includes a memory (a storage unit 260 described later). In the memory, an individual ID for specifying each conversation collection device 200 is stored so that the wearer can be distinguished. In the minutes creation support system, by linking the individual ID with, for example, an in-house personnel HR DB (not shown), the voice data detected by the microphone 201 of each conversation collection device 200 is distinguished from the voice data. can do.

  The infrared transmitting / receiving unit 202 encodes the individual ID stored in the memory and transmits it by infrared. Further, the infrared transmission / reception unit 202 can receive the individual ID transmitted by the infrared transmission / reception unit 202 of another conversation collection device 200.

  In the minutes creation support system, when another individual ID is received by the conversation collection device 200, a participant wearing the conversation collection device 200 wears the conversation collection device 200 that has received the individual ID. It is detected that it is meeting with a participant who is present.

FIG. 4 is a diagram schematically illustrating an example of a hardware configuration of the conversation collection device 200.
Referring to FIG. 4, a conversation collection device 200 includes a CPU (Central Processing Unit) 251, a RAM (Random Access Memory) 252, and a ROM (Read) configured by an arithmetic device that controls the operation of the terminal as a whole. Only Memory) 253, a communication device 254, and a storage unit 260.

  The RAM 252 functions as a work area when the CPU 251 executes a program. The communication device 254 includes a network communication device such as a LAN (Local Area Network) card or a LAN adapter. Note that when data is transmitted to the server 100 during a meeting or the like, it is desirable to have a non-contact communication function such as a wireless LAN card or a Bluetooth (registered trademark) chipset. The conversation collection device 200 is connected to the server 100 via the communication device 254.

  The storage unit 260 includes a program storage unit 261 that stores programs executed by the CPU 251 and a data storage unit 262 that stores various data. Note that the storage unit 260 may be configured by a recording medium that is detachable from the main body of the conversation collection device 200. Storage media include CD-ROM (Compact Disc-Read Only Memory), DVD-ROM (Digital Versatile Disk-Read Only Memory), USB (Universal Serial Bus) memory, memory card, FD (Flexible Disk), hard disk, magnetic Tape, cassette tape, MO (Magnetic Optical Disc), MD (Mini Disc), IC (Integrated Circuit) card (excluding memory cards), optical card, mask ROM, EPROM, EEPROM (Electronically Erasable Programmable Read-Only Memory), etc. And a medium for storing the program in a nonvolatile manner.

Server configuration
FIG. 5 is a diagram schematically illustrating an example of the hardware configuration of the server 100.

  Referring to FIG. 5, the server 100 includes a CPU 151, a RAM 152, a ROM 153, a communication device 154, a display driver 155, and a display device 156 that are configured by an arithmetic device that generally controls the operation of the server 100. And an input unit 157 and a storage unit 160.

  The communication device 154 includes a network communication device such as a LAN card or a LAN adapter. Note that when receiving data from the conversation collection device 200 during a meeting or the like, it is desirable to have a non-contact communication function such as a wireless LAN card or a Bluetooth (registered trademark) chipset. The server 100 communicates with the conversation collection device 200 via the communication device 154.

  The display device 156 includes a liquid crystal display device. The display driver 155 generates image information to be displayed on the display device 156 based on information generated by the CPU 151 executing an appropriate program. The display driver 155 may be provided as hardware as illustrated in FIG. 5, or may be provided as software by the CPU 151 executing an appropriate program.

  The input unit 157 is realized by an input device such as a touch sensor, a keyboard, and a mouse, for example.

  The storage unit 160 includes a program storage unit 161 that stores a program executed by the CPU 151, and a data storage unit 162 that stores various data required for the server 100 to realize the functions described in the present embodiment. Including.

  In the minutes creation support system, each component of the tag creation unit 910, the data storage unit 920, and the minutes creation system 800 of the life log creation system 900 shown by a broken line in FIG. The The tag generation unit 910, the search unit 810, the specific speaker identification unit 830, the conversation concentration calculation unit 840, the extraction unit 850, the priority setting unit 860, the link generation unit 870, and the minutes creation unit 880 are, for example, the CPU 151. This is realized by executing an appropriate program. The data storage unit 920 is realized by the storage unit 260, for example. The input unit 820 is realized by the input unit 157 and the communication device 154, for example.

[Conversation detection]
Next, how the establishment of a conversation between participants is detected in the minutes creation support system will be described.

FIG. 6 is a diagram schematically illustrating an example of a meeting scene.
Referring to FIG. 6, six participants 300 </ b> A to 300 </ b> F participate in the conference. In addition, FIG. 6 is equivalent to the figure which looked at six participants 300A-300F sitting around a long desk from the top. The participants 300A to 300F wear their own conversation collection devices 200 (conversation collection devices 200A to 200F). In FIG. 6, the conversation collection devices 200 </ b> A to 200 </ b> F are shown larger than the participants 300 </ b> A to 300 </ b> F for easy understanding.

  The infrared transmission / reception unit 202 of the conversation collection device 200 has high directivity. The infrared transmission / reception unit 202 of each conversation collection device 200 can communicate with the infrared transmission / reception unit 202 of the other conversation collection device 200 only when facing the infrared transmission / reception unit 202 of the other conversation collection device 200. .

  In FIG. 6, when the participant 300A and the participant 300D face each other, the conversation collection device 200A and the conversation collection device 200D face each other, so that the infrared transmission / reception unit 202A and the infrared transmission / reception unit 202D are indicated by an arrow R1. Communication is possible along the route. The conversation collection device 200A transmits to the data storage unit 920 together with the received time information that the infrared transmission / reception unit 202A has received the individual ID of the infrared transmission / reception unit 202D. In addition, the conversation collection device 200D transmits to the data storage unit 920 together with the received time information that the infrared transmission / reception unit 202D has received the individual ID of the infrared transmission / reception unit 202A.

  The conversation concentration degree calculation unit 840 determines whether there is a set of conversation collection devices 200 that have received the individual IDs of the other party at the same time. If it is determined that such a group exists, it is determined that a conversation is established between the participants corresponding to both of the groups. That is, for example, according to the above detection example, the conversation concentration degree calculation unit 840 establishes a conversation between a participant corresponding to the conversation collection device 200A and a participant corresponding to the conversation collection device 200D. Judge.

  As described above, the conversation concentration degree calculation unit 840 can detect when and among the participants the conversation is established. Thereby, the conversation concentration degree calculation part 840 can generate | occur | produce the information showing the time change of a participant's conversation state in a meeting. FIG. 7 is a diagram schematically illustrating an example of a conversation state that is generated.

In the conversation state of FIG. 7, the conversation (speech) part of ST1 to ST10 is shown.
In the conference shown in FIG. 7, four participants A, B, C, and D are shown. ST1 shows a state in which a conversation is established for a short time between Mr. A and each of Mr. B, Mr. C, and Mr. D.

  ST2 shows a state in which a conversation is established between Mr. A and Mr. B, and Mr. B is talking to Mr. A. Which party is talking to which participant among the participants having a conversation is determined based on which microphone 201 of both of them has acquired the voice. That is, FIG. 7 shows that sound is detected in the microphone 201 of the conversation collection device 200 worn by Mr. B during the period of ST2.

  Note that the conversation concentration degree calculation unit 840 uses the filter of a predetermined volume for the voice data acquired by the microphone 201 to separate the environmental sound from the acquired voice, and wears the conversation collection device 200. It is preferable to determine whether or not the person has made a voice. Further, the conversation concentration degree calculation unit 840 acquires voice data of a person wearing each conversation collection device 200 as a sample, and a period during which the person speaks only during a period in which voice data that can be identified as the voice of the person is detected. You may judge.

  ST3 shows a state in which a conversation is established between Mr. A and Mr. B, and Mr. A is talking to Mr. B. ST4 and ST5 indicate a state in which a conversation is established between Mr. C and Mr. D, and a statement from Mr. C to Mr. D and a statement from Mr. D to Mr. C are frequently performed alternately. ST6 and ST7 show a state in which a conversation is established between Mr. B and Mr. D, and a statement from Mr. B to Mr. D and a statement from Mr. D to Mr. B are alternately performed frequently. ST8 shows a state in which a conversation is established between Mr. A and Mr. C, and Mr. C is talking to Mr. A. ST9 shows a state in which a conversation is established between Mr. A and Mr. C and Mr. A is talking to Mr. C. ST10 shows a state in which a conversation is established between Mr. A and Mr. B, and Mr. B is speaking to Mr. A.

  Based on the information acquired by each conversation collection device 200 and stored in the data storage unit 920, the conversation concentration degree calculation unit 840 includes the conversation state as illustrated in FIG. The conversation time of each participant can be calculated.

FIG. 8 is a diagram schematically illustrating an example of the conversation time of each participant that is generated.
FIG. 8 shows who, from whom to whom, how long, and how much each of Mr. A, Mr. B, Mr. C, and Mr. D made a statement. In the example shown in FIG. 8, it is assumed that the conversation time in a conference different from the conference shown in FIG. 7 is shown. FIG. 8 shows the sum of speech time of each participant for each partner for a certain conference. That is, for example, the total speech time during the conference from Mr. A to Mr. B is 3 minutes. Further, the total speech time during the conference from B to A is 2 minutes. In this specification, the conversation time means the time when conversation is established between participants. In other words, when speaking between two parties, it is the sum of the time when the speech was made and the time when the speech was received. For example, in the example of FIG. 8, the conversation time between Mr. A and Mr. B is 5 minutes. Among these, about Mr. A, the time spent speaking (to Mr. B) is 3 minutes. Also, it takes 2 minutes to speak (from Mr. B).

  Further, in this specification, the time when speech is made from one participant to the other participant may be referred to as the speech time of one participant. Furthermore, the time when speech is made from one participant to the other participant may be referred to as the talk time of the other participant.

  In the present embodiment, the conversation concentration degree calculation unit 840 is shown in FIG. 9 by specifying the conversation partner, the time of speaking, and the time of speaking for each participant of a certain meeting. Such a matrix can be created, and based on this matrix, the conversation concentration level of each participant in the conference can be calculated. FIG. 9 is a diagram for explaining a method of calculating the conversation concentration degree.

  In FIG. 9, six participants “Tanaka”, “Ito”, “Inoue”, “Ishida”, “Sato”, and “Kobayashi” are listed. In the upper table in FIG. 9, the speech time and the call arrival time of each participant are shown in minutes. In the vertical direction, the utterance time for each participant is shown. For example, for Mr. Inoue, the utterance time for Mr. Tanaka is 3 minutes, the utterance time for Mr. Ito is 0 minutes, the utterance time for Mr. Inoue himself is 0 minutes, and the utterance time for Mr. Ishida is 1 minute. Yes, the utterance time for Mr. Sato is 0 minutes, and the utterance time for Mr. Kobayashi is 0 minutes.

  The conversation concentration degree calculation unit 840 calculates the total utterance time of all participants. In the example of FIG. 9, it is 27 minutes.

  In addition, the conversation concentration degree calculation unit 840 calculates the total utterance time of all participants. In the example of Figure 9, Tanaka-san is 9 minutes, Ito-san is 3 minutes, Inoue-san is 4 minutes, Ishida-san is 7 minutes, and Sato-san is 3 minutes. Kobayashi-san is 1 minute.

  In addition, the conversation concentration degree calculation unit 840 calculates the sum total of the talk time of each participant. In the example of FIG. 9, Tanaka-san is 10 minutes, Ito-san is 3 minutes, Inoue-san is 4 minutes, Ishida-san is 5 minutes, and Sato-san is 3 minutes. Kobayashi-san is 2 minutes.

  In addition, the conversation concentration degree calculation unit 840 calculates the utterance concentration degree of each participant. The utterance concentration level is the ratio of the utterance time of each participant to the total utterance time of all participants. In the example of FIG. 9, Tanaka-san is 0.33, Ito-san is 0.11, Inoue-san is 0.15, Ishida-san is 0.26, and Sato-san is 0.11 and 0.04 for Mr. Kobayashi.

  Further, the conversation concentration level calculation unit 840 calculates the incoming call concentration level of each participant. The incoming call concentration level is the ratio of the incoming time of each participant to the total of the outgoing time of all participants. In the example of FIG. 9, Mr. Tanaka is 0.37, Mr. Ito is 0.11, Mr. Inoue is 0.15, Mr. Ishida is 0.19, Mr. Sato is 0.11 and 0.07 for Mr. Kobayashi.

  Moreover, the conversation concentration degree calculation part 840 calculates the conversation concentration degree of each participant. The conversation concentration level is the sum of the utterance concentration level and the incoming call concentration level. In the example of FIG. 9, Tanaka is 0.7. For Mr. Tanaka, the total speech time is 9 minutes and the total speech time is 10 minutes. The total utterance time for the entire conference is 27 minutes. Therefore, Tanaka's conversation concentration degree is calculated as approximately 0.7 according to “(9 + 10) / 27”. It is 0.22 for Mr. Ito, 0.3 for Mr. Inoue, 0.45 for Mr. Ishida, 0.22 for Mr. Sato, and 0.11 for Mr. Kobayashi. is there.

[Extract audio data]
In the present embodiment, specific speaker specifying unit 830 specifies a specific speaker in the conference based on the conversation concentration level of each participant. A specific speaker is considered a central participant in the conference. In the example of FIG. 9, Mr. Tanaka with the highest conversation concentration is identified as the specific speaker.

  In this embodiment, the extraction unit 850 extracts the voice data of the speech made by the specific speaker and the voice data of the speech received by the specific speaker from the voice data of the conversation stored in the data storage unit 920. Extract as important remarks at the meeting. In the example of FIG. 9, voice data of the speech that Mr. Tanaka gave to each participant and the speech received by Mr. Tanaka are extracted. FIG. 10 is a diagram schematically showing the content of the voice data corresponding to the speech extracted by the extraction unit 850 among the speeches of each person shown in FIG.

  In the meeting shown in Fig. 9, the remarks that were actually important were the remarks from Mr. Ishida to Mr. Tanaka (3 minutes in total) and the remarks from Mr. Sato to Mr. Tanaka (total). 2 minutes) and Tanaka-san's remarks to Mr. Ishida (3 minutes in total). According to the result shown in FIG. 10, all of the statements shown as important in FIG. 9 have been extracted.

[Meeting process]
FIG. 11 is a flowchart of the minutes creation process executed in the minutes creation system. The minutes creation process is realized as a minutes creation application installed in the server 100, for example.

  In the minutes creation process, first, in step SA10, the input unit 820 accepts input of information related to the meeting for creating the minutes by the user. Specifically, necessary properties such as a meeting name, a meeting date and time, a meeting place, and a meeting participant are input by the user.

  For example, as shown in FIG. 12, information related to the conference (date and time, location, participants, etc.) is registered in the server 100 as tag information. In step SA10, input unit 820 accepts input of information specifying a meeting from the registered information related to the meeting, and advances the process to step SA20.

FIG. 12 is a diagram illustrating an example of tag information registered in advance for each conference.
Referring to FIG. 12, in data storage unit 920, the conference ID, start date / time, end date / time, location, participant (participant ID), purpose, and conference form are registered as information of each conference. . The conference ID is information for identifying a conference, and a unique value is set for each conference. The start date and time is information for specifying the start time of the conference. The end date and time is information for specifying the end time of the conference. The place, the participant, and the purpose are information for specifying the meeting place, the participant, and the purpose, respectively. In particular, the participant is specified by, for example, a participant ID. In the present embodiment, each participant ID is associated with the individual ID of the conversation collection device 200 worn by each participant.

  As a meeting form, two types of information “self-assuming” and “system evaluation” are stored. “Self-assuming” is information input by an administrator or the like, for example, and “system evaluation” is evaluation content determined based on, for example, the conversation state described with reference to FIG. For example, when the conversation state has a high conversation concentration level of a certain participant and the conversation concentration levels of other participants are low to some extent, it is determined as “Q & A type”. Further, for example, when the conversation concentration level of a particular participant is particularly high and the conversation concentration levels of the remaining participants are values close to 0, it is determined as “notification notification type”. Further, for example, when the conversation concentration level of the participants is the same value, the determination of “interactive” or the like is made.

  In step SA20, the conversation concentration degree calculation unit 840 searches the life log (voice data) of the conference designated in step SA10 from the data storage unit 920, and advances the process to step SA30.

  For example, when a conference is specified, as shown in FIG. 12, the start date / time, end date / time, and participant ID corresponding to the conference are specified. For example, when the conference indicated by C1 in FIG. 12 (conference ID is “POE-0003”) is designated, the start date / time (August 27, 2011, 10:00) and the end date / time ( On August 27, 2011, 11:10), the participant IDs (10201, 10850, 10430, and 8431) are specified.

  The conversation concentration degree calculation unit 840 registers, in the data storage unit 920, all the voice data of the speech made by the participant ID or the received speech from the start date to the end date based on the specified information. Search from stored audio data.

  More specific description will be given below. FIG. 13 is a diagram illustrating an example of a screen displayed on the display device 156 of the server 100. When information specifying a conference is input, the search unit 810 displays tag information of the specified conference as shown in a screen 500 in FIG. Then, when the user (for example, the minutes creator) confirms the tag information and wishes to proceed with the creation of the minutes for the meeting, the user further operates the link creation button 501 in the screen 500. To do. In response to the operation, the search unit 810 searches for audio data specified by the tag information of the conference as described above.

  In other words, according to the process in step SA20, all audio data recorded for the attendees of the meeting becomes the search result.

  Note that the wireless LAN connected from the conference location may be specified based on the conference location in the tag information. In the LAN identified in this way, the voice data acquired (registered in the LAN server) from the start date to the end date may be used as the search result in step SA20.

  Next, in step SA30, the conversation concentration degree calculation unit 840 uses the voice data obtained as a search result in step SA20, and the conversation state as described with reference to FIGS. (Directed graph between participants) is created, and the process proceeds to step SA40.

  In step SA40, the conversation concentration degree calculation unit 840 calculates the conversation concentration degree of each participant as described with reference to FIG. 9, and advances the processing to step SA50.

  In step SA50, the specific speaker specifying unit 830 specifies a specific speaker based on the conversation concentration level of each participant calculated in step SA40, and proceeds to step SA60.

  In step SA60, the extraction unit 850 extracts voice data corresponding to the utterance made by the specific speaker and the utterance to the specific speaker from the search result of step SA20, and the process proceeds to step SA70.

  In step SA70, the link generation unit 870 generates link data for linking to each of the audio data extracted in step SA60, and the minutes creation unit 880 further creates a minutes such as the screen 500 in FIG. A minutes file is created by adding the created link data to the screen data for the meeting to be created, and the minutes creation process is terminated.

FIG. 14 is a diagram illustrating a display example of a minutes file.
Referring to FIG. 14, the minutes file includes display screen data such as screen 510. The screen 510 includes the date and time of the target meeting, location, participants, and data link. The screen 510 is a screen in which a data link is added to the screen 500 of FIG. The date, place, and participant are acquired from the tag information as shown in FIG.

  On the screen 510, a link for each group of participants who has established a conversation is displayed. The link is displayed with an underline. In addition, the display order of the group of participants in the data link on the screen 510 can be, for example, the order in which conversation time is long. That is, the screen 510 shows the case where the conversation between Mr. Tanaka and Mr. Ishida is the longest conversation time (the sum of Mr. Tanaka's speech time and call arrival time for the conversation with Mr. Ishida). Yes.

  In the data storage unit 920, the audio data is registered as a different file every time a conversation is divided. Specifically, the CPU 251 (see FIG. 4) of the conversation collection device 200 each time a silent period (or an audio period less than a specific level) of a predetermined time or more appears for the voice input from the microphone 201, Generate audio data in different files. Then, when a certain group is designated among the groups of participants displayed as data links on the screen 510, information linked to each of the voice data of the conversation of the group is displayed. FIG. 15 is a diagram illustrating an example of a screen for displaying information linked to each audio data.

  Referring to FIG. 15, on screen 520, a link to each voice data of the conversation between Tanaka and Mr. Ishida is displayed so as to include the time when the voice data was acquired. In the server 100, when each link is selected, the audio data recorded at the time corresponding to the link is reproduced. Note that the time here is, for example, the time when voice acquisition is started. Note that, on the screen 520, not the order of the time when the sound data is acquired but the properties of the corresponding sound data, such as the order in which the playback time of each sound data is long, is sorted and displayed according to a predetermined condition. .

  Returning to FIG. 14, the screen 510 includes a link indicated by the characters “other remarks”. This link corresponds to speech data of a group of utterances from the longest conversation time to a specific order in the conference. When such a link is selected, for example, the display device 156 displays a link to each audio data of the conversation of the group of participants classified as “other utterances”. When each link displayed there is selected, the server 100 reproduces audio data corresponding to the link.

  According to the minutes creation process described above, as described with reference to FIGS. 6 to 9 and the like, the state relating to the establishment of the conversation in the conference is detected, and the recorded voice data is based on the detection result. Voice data is extracted from the data and used to create the minutes.

  As a result, not all the voices acquired at the meeting but the voices considered to be important at the meeting are extracted and used to create the minutes.

[Modifications, etc.]
(Modification 1)
In the present embodiment described above, “meeting” refers to an object for which minutes are to be created, and is not limited to an event that is scheduled to be gathered specially. For example, in an office, information for identifying a person who is assumed to be in the office is specified as a participant ID, and each user in the office wears the conversation collection device 200 described with reference to FIG. And even when the daily conversation in the office is recorded, the technical idea disclosed by this embodiment can be applied.

(Modification 2)
The conversation collection device 200 constitutes an example of detection means. The conversation collection device is not limited to the form described with reference to FIG. FIG. 16 is a diagram illustrating an appearance of a modified example of the conversation collection device. FIG. 17 is a diagram schematically illustrating an example of a hardware configuration of the conversation collection device in FIG. 16.

  With reference to FIGS. 16 and 17, the conversation collection device 210 includes a microphone 211 and a camera 212 instead of the microphone 201 and the infrared transmission / reception unit 202 in the configuration described with reference to FIG. 4.

  In the conversation collection device 210, the camera 212 captures an image existing ahead of the line of sight of the participant of the conference wearing the conversation collection device 210. Therefore, the conversation concentration degree calculation unit 840 identifies who the participant wearing the conversation collection device 210 is looking at by analyzing an image captured by the camera 212 according to a known face discrimination algorithm. be able to.

  Furthermore, the conversation concentration degree calculation unit 840 can determine whether the participant wearing the conversation collection device 210 is speaking by analyzing the image together with the face determination algorithm. By further using the discrimination result, the participant wearing the conversation collection device 210 based on the known nod theory (Okayama Prefectural University, Department of Information Systems Engineering, Professor Toshio Watanabe's “Physical Retraction Technology iRT”, etc.) It is possible to identify the other party's conversation partner with higher accuracy, for example, by identifying the other party who is watching while speaking.

  Moreover, in the conversation collection device 210, the camera 212 can photograph the other party reflected on the television screen. Therefore, the information collected by the conversation collection device 210 can be used for specifying a conversation partner even in a video conference with a remote place.

(Modification 3)
FIG. 18 is a diagram schematically illustrating an example of a hardware configuration of another modified example of the conversation collection device 200.

  In this modification, the conversation collection device 200 includes a biological information detection unit 203 in addition to the configuration shown in FIG. The biological information detection unit 203 is a sensor for detecting biological information (vital data) such as the heart rate, heat, and / or skin conductivity of the user wearing the conversation collection device 200. The participant who wears the conversation collection device 200 further wears the sensor in such a manner that data to be acquired by the sensor can be acquired.

  In the present modification, the data storage unit 920 also records the biometric information of the participant who has issued the audio data and / or the participant who has received the speech of the audio data, along with the audio data.

  The biometric information recorded in this way is used to specify the content and importance of the conversation, such as how the speech corresponding to the audio data was uttered and how much impact was given to the conversation partner. It can be supporting information. In particular, there is also a study (“normal / non-normal (sleepiness, tension) estimation method using pulse rate” NTT) that identifies presentation and question intervals from the pattern of heartbeat interval time-series data. Therefore, the accuracy of detecting the conversation state can be improved by recording the biological information as described above.

  Furthermore, the biometric information recorded as described above can be used not only for detecting the conversation state but also for specifying the importance level of the conversation. For example, when the heart rate of a speaker when voice data is acquired is higher than a specific rate with respect to the normal heart rate of the speaker, the speech is considered to be in a tense state. . As can be seen, a link of the audio data may be displayed. FIG. 19 is a diagram showing a modification of the link display screen of FIG. In the screen 530 shown in FIG. 19, a 10:50 message link and a 10:44 message link are displayed with a heart symbol. The heart mark indicates that the participant who made the speech made the corresponding speech in a tension state as described above.

  It can be said that there is a relatively high possibility that an utterance in tension is an important content. Users who play the minutes file grasp the contents of the meeting earlier by giving priority to playing audio files that are relatively likely to be important contents according to the display of the heart symbol. It becomes possible.

(Modification 4)
FIG. 20 is a diagram schematically illustrating an example of a hardware configuration of still another modified example of the conversation collection device 200. The conversation collection device 220 of FIG. 20 includes an electronic compass 204 and a GPS (Global Positioning System) controller 205 instead of the infrared transmission / reception unit 202 with respect to the conversation collection device 200 described with reference to FIG.

  The electronic compass 204 includes a magnetic sensor and detects the orientation of the conversation collection device 220. The GPS controller 205 acquires position information of the conversation collection device 220 by receiving a GPS signal or a position signal from the base station.

  In this modification, the data storage unit 920 registers the position information and orientation information of the conversation collection device 220 (instead of the individual ID acquired by the infrared transmission / reception unit 202). Thus, in the minutes creation support system, the data storage unit 920 acquires the position information and orientation information of each participant (the user wearing the conversation collection device 220). Thereby, the position and orientation of each user can be specified. Thereby, it is possible to identify a set of users that are located in the vicinity (within a predetermined distance range) and face each other. Then, the conversation concentration degree calculation unit 840 establishes a conversation between the users of the set on the condition that the voice is detected in the microphone 201 of the conversation collection device 220 worn by one user in the set. Can be identified.

(Modification 5)
In the present embodiment described above, weighting is performed by setting a weight value for each participant, and the specific speaker specifying unit 830 can further specify the specific speaker by further using the weighting. it can.

  For example, personnel information in the workplace is used as the weight value assigned by weighting. In other words, people in managerial positions often make decisions at meetings because they have many decisions in the workplace. For example, it is assumed that a participant in the position of general manager has a weight value twice that of other participants. The specific speaker specifying unit 830 calculates the product of the conversation concentration level of each participant and the weight value assigned to each participant, and determines the participant having the largest value as the specific speaker. can do.

  Further, the weight value assigned to each participant may be changed for each conference according to the agenda of the conference.

(Modification 6)
The value of the weight used when the conversation concentration degree calculation unit 840 identifies a specific speaker may be set according to the time zone when the speech occurs. For example, in a two-hour meeting, the utterances in the latter half hour may be more important than the utterances in the first half hour due to reasons such as the meeting place being mature. Therefore, when calculating the speech time and speech time as described with reference to FIG. 9, for example, the length of the speech time and speech time for speech in the latter half 1 hour is multiplied by 1.5, It may be used to calculate the conversation concentration level. In this case, the length of the talk time and the talk time for the speech in the first half hour is left as the actual time.

  As a result, it is possible to specify the specific speaker by weighting the voice data according to the time zone in which the speech has occurred. Moreover, the priority at the time of using the audio | voice of the said participant for the minutes can be attached to the participant who made the speech according to the time zone when the speech occurred.

(Other variations)
Opinion exchange-type meetings such as brainstorming are rarely made of minutes, and ideas that come out on the spot are often gathered on the spot. Therefore, it is considered to refer mainly to images on the whiteboard used on the spot and information at the time when additional information (post-it) was added during the conference.

  However, there is no facilitator even in an opinion exchange meeting, and a specific speaker (main speaker) can be found. Eventually, there are cases where summaries are made by facilitators in opinion exchange meetings. Therefore, linking the speech of the main speaker and the voice made to the main speaker is considered to be useful in an opinion exchange type meeting.

  Note that the minutes creation support system described in this specification can create a minutes file during a meeting or after a meeting. If it is during a meeting, a link will be produced | generated from the conversation until then, and a link will be updated by a conversation progressing sequentially.

  The minutes creation support system described in this specification extracts information (voice data) important for creating minutes from the conversation concentration level, and creates a link.

  Thereby, the user who refers to the minutes file can refer to the original data of the minutes without going through all the information (voice data) recorded in the meeting.

  Note that the degree of conversation concentration acquired in the minutes creation support system can be used for the following purposes.

・ Comparing each person's concentration of conversation and checking whether or not discussions are possible between participants ・ Extracting important comments from the concentration during conversation and applying security (changing reading attributes and encryption) Facilitator skill check set before the meeting (speaking by yourself or activating discussion)
・ Remarks that are not related to the main speaker are unnecessary as a record, so they can be deleted or obfuscated (encryption, data division, etc.), data compression, chatting unrelated to discussion, etc., and / or Or protection of privacy It should be thought that each embodiment indicated this time is an illustration and restrictive at no points. The scope of the present invention is defined by the terms of the claims, rather than the description above, and is intended to include any modifications within the scope and meaning equivalent to the terms of the claims. In addition, the invention described in the embodiment and each modification is intended to be carried out independently or in combination as much as possible.

  100 server, 200, 200A to 200F conversation collection device, 201, 211 microphone, 202 infrared transmission / reception unit, 203 biological information detection unit, 204 electronic compass, 205 GPS controller, 212 camera, 800 minutes creation system, 900 life log generation system .

Claims (13)

Detection means for detecting a state in which a conversation between the participants of the conference is established;
Recording means for recording the voices uttered by the participants of the conference;
An extracting means for extracting a specific sound from the sound recorded in the recording means based on the detection result of the conversation establishment state by the detecting means;
A minutes creation means for creating minutes of the meeting using the specific voice ;
The detection means is configured so that each of the infrared transmission / reception units of the devices having individual IDs worn by each participant can communicate with each other, and there is a set of the participants who have received the individual IDs of the other party. of conversation among participants that be determined that the state has been established, the minutes creating support device. Detection means for detecting a state in which a conversation between the participants of the conference is established;
Recording means for recording the voices uttered by the participants of the conference;
An extracting means for extracting a specific sound from the sound recorded in the recording means based on the detection result of the conversation establishment state by the detecting means;
A minutes creation means for creating minutes of the meeting using the specific voice;
The detection means specifies the position and orientation of each participant by means of an electronic compass of a device worn by each participant and a GPS controller, and the sound is heard in a pair of participants located in the vicinity and facing each other. A minutes creation support apparatus that, when detected, determines that a conversation between participants in the conference is established. The extraction means includes
For each participant of the conference, calculate the total conversation time by accumulating the time determined by the detection means that the conversation was established,
Based on the total conversation time for each participant, identify the main speaker in the meeting,
The minutes creation support apparatus according to claim 1, wherein a voice of a conversation including the main speaker is extracted as the specific voice from the voice recorded by the recording unit. The extraction means includes
Calculate the conversation concentration of each participant by calculating the ratio of each participant's total conversation time to the total conversation time for all participants;
Based on the conversation concentration of each participant, the participant having the highest conversation concentration is determined as the main speaker,
4. The minutes creation support apparatus according to claim 3 , wherein a voice of a conversation including the main speaker is extracted as the specific voice from the voice recorded by the recording means. It said recording means is to distinguish between voice each participant emitted voice and the participant that originated, recording audio, minutes creation support according to any one of claims 1 to 4 apparatus. The proceedings preparation means, said generating a link to a particular voice, presenting the link, minutes creation support apparatus according to any one of claims 1 to 5. The extraction means based on the priority based on the time zone of the voice utterance, the extracted specific voice from the recorded voice in the recording means, according to any one of claims 1 to 6 Minutes creation support device. The extraction means extracts the specific sound from the sound recorded in the recording means based on a priority set for a participant who has emitted the sound recorded in the recording means. minutes creation support apparatus according to any one of-claims 7. Said recording means, said speech, recorded together with biometric information of a person who emits the sound, minutes creation support apparatus according to any one of claims 1 to 8. A minutes creation support system comprising detection means worn by a participant of a meeting and a minutes creation support device capable of communicating with the detection means,
The minutes creation support device,
Based on the detection result obtained from each of the detection means, detection means for detecting a state in which the conversation between the conference participants is established,
Recording means for recording the voices uttered by the participants of the conference;
An extracting means for extracting a specific sound from the sound recorded in the recording means based on the detection result of the conversation establishment state by the detecting means;
Only contains the proceedings creating means for creating the minutes of the meeting by using the specific voice,
The detection means is configured so that each of the infrared transmission / reception units of the devices having individual IDs worn by each participant can communicate with each other, and there is a set of the participants who have received the individual IDs of each other. of conversation among participants that be determined that the state has been established, the minutes creation support system. A minutes creation support system comprising detection means worn by a participant of a meeting and a minutes creation support device capable of communicating with the detection means,
The minutes creation support device,
Based on the detection result obtained from each of the detection means, detection means for detecting a state in which the conversation between the conference participants is established,
Recording means for recording the voices uttered by the participants of the conference;
An extracting means for extracting a specific sound from the sound recorded in the recording means based on the detection result of the conversation establishment state by the detecting means;
A minutes creation means for creating the minutes of the meeting using the specific voice;
The detection means specifies the position and orientation of each participant by means of an electronic compass of a device worn by each participant and a GPS controller, and the sound is heard in a pair of participants located in the vicinity and facing each other. A minutes creation support system that, when detected, determines that a conversation between participants in the conference is established. A computer readable program comprising:
Detection means for detecting the state in which the conversation between the participants of the conference is established,
A recording means for recording a voice uttered by a participant of the conference;
An extracting means for extracting a specific sound from the sound recorded in the recording means based on the detection result of the conversation establishment state by the detecting means; and
Function as a minutes creation means for creating the minutes of the meeting using the specific voice ,
The detection means is configured so that each of the infrared transmission / reception units of the devices having individual IDs worn by each participant can communicate with each other, and there is a set of the participants who have received the individual IDs of the other party. Minutes creation program that judges that a conversation between participants has been established . A computer readable program comprising:
Detection means for detecting the state in which the conversation between the participants of the conference is established,
A recording means for recording a voice uttered by a participant of the conference;
An extracting means for extracting a specific sound from the sound recorded in the recording means based on the detection result of the conversation establishment state by the detecting means; and
Function as a minutes creation means for creating the minutes of the meeting using the specific voice,
The detection means specifies the position and orientation of each participant by means of an electronic compass of a device worn by each participant and a GPS controller, and the sound is heard in a pair of participants located in the vicinity and facing each other. A minutes creation program that, when detected, determines that a conversation is established between participants in the conference.

Images