KR102548365B1 - Method for generating conference record automatically and apparatus thereof - Google Patents

Abstract

According to an embodiment of the present invention, a method of automatically generating a meeting record performed by an apparatus for automatically generating a meeting record includes receiving contents of a plurality of speakers participating in a conference, converting the received contents into text for each speaker, , Extracting a keyword from the converted text, determining a first keyword based on the keyword extracted from the converted text, and generating comment text for each speaker based on the determined first keyword. and generating a meeting record for the first keyword including the generated comment text for each speaker.

Description

Meeting record automatic generation method and its device

The present invention relates to a method and apparatus for automatically generating a meeting record. More particularly, it relates to a method and apparatus for automatically generating a meeting record by subject by identifying a subject of an ongoing meeting.

With the development of multimedia environments and communication technologies, services and devices capable of maximizing time and space efficiency in conducting conferences are being provided.

Using these services and devices, a vast amount of content required for a conference can be displayed and shared with participants, and the participants' writings on the conference content can also be converted into data and stored. In addition, participants may be provided with a real time conference service in different spaces.

At this time, if you attend the meeting while it is in progress or check the contents of the meeting after the meeting is over, it is inefficient to check the record of the entire meeting again. Selective services are required.

Nevertheless, if it is necessary to check the contents of a meeting discussed at a specific time, the contents of a meeting only related to a specific topic, or the comments of a specific speaker, a service is provided to create and provide personalized meeting minutes reflecting the meeting record requested by the requester. It's not happening.

Korean Patent Publication No. 10-2015-0117796 (published on October 21, 2015)

A technical problem to be solved by the present invention is a method and apparatus for allowing a person who does not attend a meeting or a participant who is present in the middle of a meeting to access a desired portion of the meeting record after the meeting is over is to provide

Specifically, the technical problem to be solved by the present invention is a method for providing information desired by a requestor for a meeting record, such as discussion contents at a specific time, meeting contents related only to a specific topic, or comments of a specific speaker, to the terminal of the meeting record requester. is to provide

Another technical problem to be solved by the present invention is to provide a method and apparatus for identifying a conference topic in conference content displayed during a conference and generating a conference record for each identified conference topic.

Another technical problem to be solved by the present invention is to provide a method and apparatus capable of collecting and converting content of conference participants into text and reflecting it in meeting minutes.

The technical problems of the present invention are not limited to the technical problems mentioned above, and other technical problems not mentioned will be clearly understood by those skilled in the art from the description below.

In order to solve the above technical problem, an aspect of a method for automatically generating a meeting record performed by an apparatus for automatically generating a meeting record includes receiving contents of a plurality of speakers participating in a conference, and converting the received contents into text for each speaker. , extracting a keyword from the converted text, determining a first keyword based on the keyword extracted from the converted text, and generating the comment text for each speaker based on the determined first keyword. and generating a meeting record for the first keyword including the generated comment text for each speaker.

In addition, another aspect of a method for automatically generating a meeting record performed by an apparatus for automatically generating a meeting record for solving the above technical problem is the step of receiving contents of a plurality of speakers participating in a conference, and the received contents for each speaker. Converting to text, receiving keywords from a user terminal, extracting keywords matching the received keywords from the converted text, and generating comment text for each speaker based on the extracted keywords and generating a meeting record for the received keyword, including the generated comment text for each speaker.

In addition, to solve the technical problem, a network interface communicating with one or more processors and a terminal of a speaker attending a conference, a memory for loading a computer program executed by the processor, and storing the computer program, The computer program includes an operation of receiving content of a plurality of speakers attending a conference, an operation of converting the received content into text for each speaker, and an operation of extracting a keyword from the converted text, and an operation of extracting a keyword from the converted text. An operation for determining a first keyword based on a keyword, an operation for generating a comment text for each speaker based on the determined first keyword, and a conference record for the first keyword including the generated comment text for each speaker. may contain an operation that creates

According to the present invention, even if a participant attends a meeting in the middle of a meeting, there is an effect of being provided with a meeting minutes generation device that enables a participant to quickly grasp the contents of a meeting before attending.

In addition, according to the present invention, there is an effect of providing a method and apparatus capable of generating a meeting record for each subject and searching for the meeting record for each subject using a keyword.

1 is a configuration diagram of a system for automatically generating a meeting record according to an embodiment of the present invention.
2 is a block diagram of an apparatus for automatically generating a meeting record according to another embodiment of the present invention.
3 is a flowchart of a method for automatically generating a meeting record according to another embodiment of the present invention.
4 is an example of content for each subject according to the order of conference progress, which is referred to in some embodiments of the present invention.
5 is an exemplary diagram for explaining a process of converting audio content of a conference participant into text and generating comment text, which is referred to in some embodiments of the present invention.
6 is an exemplary diagram for explaining a method of identifying a keyword for a topic of an ongoing conference according to another embodiment of the present invention.
7 is an example of an automatically generated conference record, referenced in some embodiments of the present invention.

Hereinafter, preferred embodiments of the present invention will be described in detail with reference to the accompanying drawings. Advantages and features of the present invention, and methods for achieving them, will become clear with reference to the embodiments described below in detail in conjunction with the accompanying drawings. However, the present invention is not limited to the embodiments disclosed below and may be implemented in various different forms, only the present embodiments make the disclosure of the present invention complete, and the common knowledge in the art to which the present invention belongs It is provided to fully inform the holder of the scope of the invention, and the present invention is only defined by the scope of the claims. Like reference numbers designate like elements throughout the specification.

Unless otherwise defined, all terms (including technical and scientific terms) used in this specification may be used in a meaning commonly understood by those of ordinary skill in the art to which the present invention belongs. In addition, terms defined in commonly used dictionaries are not interpreted ideally or excessively unless explicitly specifically defined. Terminology used herein is for describing the embodiments and is not intended to limit the present invention. In this specification, singular forms also include plural forms unless specifically stated otherwise in a phrase.

In this specification, content refers to various types of information that a conference participant provides to an automatic conference record creation system. For example, the content may include audio content that is a speech of a conference participant.

Also, the content may include text content provided by conference participants. In this case, the text content may be input from a conference participant through a terminal of the conference participant, and includes text content input through various input units of the terminal of the conference participant.

For example, it is assumed that an apparatus for automatically generating conference records according to an embodiment of the present invention provides a real-time conference service to terminals of participants participating in a conference in a separate space. At this time, conference participants can participate in the conference by transmitting and receiving various contents through various input units of the terminal. In this case, the text content may include text input through the keyboard of the terminal. For example, this is a case where a conference is conducted by chatting between conference participants. Also, the text content may include text input in a handwritten form through a note-taking device of a terminal. For example, this is a case in which handwritten input of a conference participant is input through a note-taking device on various documents displayed on a terminal of a conference participant.

Hereinafter, a case in which content according to an embodiment of the present invention is audio content is mainly described, but the embodiment of the present invention is not limited thereto.

1 is a configuration diagram of a system for automatically generating a meeting record according to an embodiment of the present invention.

Referring to FIG. 1 , the system for automatically generating a meeting record may include an apparatus for automatically generating a meeting record 100 and a display device 200 . Also, although not shown, the automatic meeting record generation system may include a separate user terminal for reading the meeting record created through the automatic meeting record generation device 100 . Through the user terminal, a user who could not attend the meeting or who participated in the middle of the meeting after the meeting was in progress can view the meeting record 300 created through the automatic meeting record generation device 100 through the user terminal. In particular, in FIG. 1 , a case in which three participants, speaker A, speaker B, and speaker C, are attending a conference is shown as an example.

The automatic meeting recording generating device 100 is a computing device connected to the display device 200 by wire or wirelessly and capable of communicating with an external device such as the user terminal.

The apparatus 100 for automatically generating a conference record may receive voice content from a speaker participating in a conference. To this end, the device 100 for automatically generating meeting records may include an input device for collecting voice content. For example, the input device may be a microphone. In this case, when a conference is held at the same place, the speeches of speakers A, B, and C, who are participants in the conference, are input to the apparatus 100 for automatically generating a conference record as audio content.

Meanwhile, the automatically generating conference record 100 may provide a real-time conference service to terminals of participants participating in a conference in a separate space. In this case, the apparatus 100 for automatically generating a conference record may receive the audio content collected by the terminals of the attendees (eg, speaker A, speaker B, and speaker C) through a network. may be The participant's terminal may be a terminal device including an output device for displaying conference contents including conference documents and participant information to participants and an input device for collecting audio and video of the participant in order to receive a real-time conference service. there is. In addition, the terminal device may include a communication device for communicating with the automatic meeting record creation device 100 . When speakers A, speakers B, and speakers C, who are participants, participate in a conference and speak through their respective terminals, voice content received through the input device of each terminal is transmitted to the conference recording generating device 100 through a network.

In another embodiment of the present invention, the system for automatically generating conference records may include a separate real-time conference service providing server. In this case, the conference content may be generated in the real-time conference service providing server and transmitted to each participant's terminal. In addition, the audio contents of the attendees may be received by the real-time conference service providing server and transmitted to the automatic meeting record creation device 100 .

The apparatus 100 for automatically generating a meeting record may convert the received voice content into text. In this case, the conference recording generating device 100 may convert audio contents of speakers A, speakers B, and speakers C who participated in the conference into text, respectively. That is, the audio contents of speaker A are converted into text of speaker A, and the audio contents of other speakers B and C are also converted into texts of speakers B and C, respectively. To this end, the meeting record generation device 100 may include software for executing a speech to text (STT) function widely known in the art to which the present invention belongs.

The apparatus 100 for automatically generating a meeting record may extract keywords from the converted text for each speaker. In this case, the conference recording generating apparatus 100 may extract keywords for each text of speaker A, speaker B, and speaker C, respectively. For example, the apparatus 100 for automatically generating a meeting record may extract keyword 1, keyword 2, and keyword 3 from speaker A's text, and extract keyword 4 and keyword 5 from speaker B's text. Also, keyword 6, keyword 7, and keyword 8 may be extracted from the text of speaker C. To this end, the automatic meeting record generation device 100 may include software for executing a keyword extraction algorithm widely known in the art to which the present invention belongs. A specific keyword extraction method will be described later in the description of FIGS. 5 and 6 .

The apparatus 100 for automatically generating meeting records may classify, extract, and store the extracted keywords for each speaker. Hereinafter, for convenience of description, a case in which participants attending a conference are speakers A and speakers B will be described as an example.

The keywords extracted from the one speaker are different keywords, but may be keywords that are related to each other. For example, keyword 1, keyword 2, and keyword 3 of speaker A may be “smart phone,” “mobile phone,” and “mobile terminal,” respectively.

Also, keywords 4 and 5 of another speaker, speaker B, may be "iPhone" and "Galaxy S", respectively.

The apparatus 100 for automatically generating a meeting record may identify a relationship between keyword 1, keyword 2, and keyword 3. To this end, the apparatus 100 for automatically generating meeting records may include a dictionary for recognizing keywords and semantic relationships between keywords. The apparatus 100 for automatically generating a meeting record may obtain additional information by searching for a keyword through a network and reflect it in advance.

As in the above example, it is assumed that the automatic meeting recording device 100 extracts “iPhone” and “Galaxy S” as keywords 4 and 5, respectively. In this case, if the "iPhone" or "Galaxy S" is not searched in the dictionary, the device for automatically creating a meeting record 100 accesses a web page through a network, and keywords 4 and 5 are specific brand names of smart phones. can be retrieved and saved.

Accordingly, the apparatus 100 for automatically generating a meeting record can recognize a semantic relationship between keyword 1, "smart phone," keyword 4, "iPhone," and keyword 5, "Galaxy S."

The apparatus 100 for automatically generating a meeting record may determine a first keyword based on the extracted keyword. Here, the first keyword may be a keyword for a topic of an ongoing conference. The apparatus 100 for automatically generating a meeting record may determine a keyword of a conference topic being discussed by speakers A and B based on keywords in text extracted from the utterances of speakers A and B. In the above example, the automatically generating conference record device 100 may recognize the semantic relationship between the extracted keywords 1 to 5, and determine the keyword of the ongoing conference topic as “smart phone”.

The apparatus 100 for automatically generating a conference record may generate comment text for each speaker based on the determined keyword of the conference topic. The comment text for each speaker may be a text in which remarks for each speaker are summarized. That is, the apparatus 100 for automatically generating a conference record may obtain text for each speaker by converting audio content for each speaker, and generate comment text for each speaker based on the obtained text for each speaker.

For example, when the utterances are made in the order of the first utterance of speaker A, the first utterance of speaker B, the second utterance of speaker A, and the second utterance of speaker B, the apparatus 100 for automatically generating a meeting record is A comment text of speaker A may be generated based on the first utterance and the second utterance. Also, the apparatus 100 for automatically generating a meeting record may generate comment text of speaker B based on the first and second remarks of speaker B. That is, the apparatus 100 for automatically generating a conference record may generate comment text for each speaker by classifying speakers and combining utterances of each speaker in a time-sequential manner.

In this case, the utterances of speaker A and speaker B may be generated based on the determined first keyword. That is, in the above example, among speaker A's first utterance and second utterance, when "smart phone" is included in the first utterance and not included in the second utterance, the automatic meeting record generation device 100 automatically generates the conference record of speaker A in the second utterance. 1 Comment text can be created by assigning weight to utterances. In this case, the first utterance may be reflected in a higher proportion than the second utterance in the comment text of speaker A. The weight and comment text reflection ratio may be set in advance by a manager of the automatic meeting record creation device 100 .

The apparatus 100 for automatically generating a meeting record may generate the conference record 300 including the generated comment text for each speaker. In the above example, the conference record generation device 100 may create a record 300 for "smart phone", which is a keyword of a conference subject in an ongoing conference. The conference record 300 may include comment texts of speakers A and speakers B who participated in the conference on “smart phone”. Additionally, the meeting record 300 may include content on a meeting document displayed through the display device 200 during a meeting on a “smart phone”.

The conference recording generating device 100 may identify a conference document displayed on the display device 200 while a conference on the “smart phone” is in progress, and extract contents from the identified conference document. The conference recording generation device 100 may identify a conference document displayed while a conference is in progress, and extract the displayed page as the content. Alternatively, the meeting record generating device 100 may identify a paragraph including “smart phone” as the determined first keyword on the conference document and extract the paragraph.

While the conference is in progress, the display device 200 may display content received from the automatically generating conference record device 100 or a separate terminal.

In particular, the display device 200 is a device that displays the meeting document while the meeting is in progress. The display device 200 may be a component of the automatic meeting record generating device 100 or may be a separate device.

The display device 200 may display various conference documents used in the conference process. For example, the display device 200 displays a document to be displayed according to the input of the conference moderator in a situation in which a conference is in progress based on document 1 for “smart phone,” which is the meeting subject, in document 1 for “smart phone.” It can be displayed by changing it to another document, Document 2. Also, the display apparatus 200 may receive a display request for document 1 again and display document 1 again. In this case, the apparatus 100 for automatically generating a meeting record may identify document 1, document 2, and document 1, which are the order of documents displayed during the meeting for “smart phone,” which is the keyword of the conference topic. Also, the apparatus 100 for automatically generating a meeting record may identify a displayed page of each document and/or a paragraph including “smart phone”.

Meanwhile, the display device 200 may be excluded from the automatic meeting record generation system when participants participate in a meeting through a real-time conference service in a separate space. In this case, a conference document transmitted from the automatically generating conference record 100 or the real-time conference service providing server may be displayed on each terminal of the participant. The apparatus 100 for automatically generating a meeting record may identify at least one of the meeting document being transmitted, a page of the meeting document, and a paragraph including the first keyword for the meeting subject.

The automatic meeting record generation device 100 may extract content related to the first keyword from all or part of the identified meeting document. Also, the automatic meeting record generation device 100 may generate the meeting record 300 in which the extracted content is reflected.

Accordingly, the automatically generating conference record 100 includes a comment text summarizing the speaker's remarks based on the keyword determined for the conference subject among keywords according to the speech between speakers and the conference content related to the determined keyword. A record 300 can be created. The generated conference record 300 may be provided to a terminal of a conference participant, and may be provided to a user who could not attend the conference or participated in the conference through the user's terminal.

Meanwhile, the meeting content on the meeting record 300 may include writing information according to a writing by a meeting participant. The writing information may be information generated according to a writing input by a participant through a terminal of a participant during a conference. Alternatively, when the display device 200 includes an input module, it may be information generated according to writing directly input into the display device 200 . The writing information will be described later in the description of FIGS. 4 and 7 .

2 is a block diagram of an apparatus 100 for automatically generating meeting records according to another embodiment of the present invention.

Hereinafter, with reference to FIG. 2 , the structure and operation of the automatic meeting record generating device 100 will be described in detail.

Referring to FIG. 2 , an apparatus for automatically generating meeting records 100 includes one or more processors 101, a network interface 102 communicating with a plurality of terminals, and a computer program executed by the processor 101 that is loaded. It may include a memory 103 and a storage 104 for storing computer programs. Also, although not shown, the device 100 for automatically generating conference records may include an input device that receives voice content of a speaker and a display device 200 that outputs conference documents.

The processor 101 controls the overall operation of each component of the automatic meeting record generating device 100 . The processor 101 may include a central processing unit (CPU), a micro processor unit (MPU), a micro controller unit (MCU), or any type of processor well known in the art. Also, the processor 101 may perform an operation for at least one application or program for executing a method according to embodiments of the present invention. The automatic meeting record generation device 100 may include one or more processors.

The network interface 102 supports wired/wireless Internet communication of the automatic meeting record creation device 100 . In addition, the network interface 102 may support various communication methods other than internet communication. To this end, the network interface 102 may include a communication module well known in the art.

The network interface 102 may perform data communication with the display device 200 shown in FIG. 1 , a terminal of an attendee, and a terminal of a user through a network. In particular, the network interface 102 may receive voice content from a participant's terminal and, if a real-time conference service providing server exists separately, receive a conference document of an ongoing conference from the real-time conference service providing server.

Memory 103 stores various data, commands and/or information. The memory 103 may load one or more programs 105 from the storage 104 to execute the method for creating a meeting record according to embodiments of the present invention. In FIG. 2 , RAM is illustrated as an example of the memory 103 .

The storage 104 may non-temporarily store the one or more programs 105 , the meeting documents database 106 , and the meeting minutes database 107 . In FIG. 2 , as an example of the one or more programs 105 , meeting recording automatic creation software 105 is illustrated.

The storage 104 may be a non-volatile memory such as read only memory (ROM), erasable programmable ROM (EPROM), electrically erasable programmable ROM (EEPROM), flash memory, or the like, a hard disk, a removable disk, or a device well known in the art. It may be configured to include any known type of computer-readable recording medium.

The automatic conference recording software 105 converts the speaker's voice content into text and extracts keywords from the converted text according to an embodiment of the present invention. In addition, the automatic meeting record creation software 105 determines a first keyword among the extracted keywords. Also, the automatic meeting record generation software 105 generates comment text for each speaker based on the first keyword and creates a meeting record.

The meeting document DB 106 stores meeting documents used for meetings. In particular, the conference document DB 106 includes pages of conference documents for each subject, conference contents displayed during the conference with respect to the first keyword, paragraphs of the conference document including the first keyword, and writings input during the conference with respect to the first keyword. Information can also be stored.

The meeting minutes database 107 may store records of each meeting created by subject. In addition, the meeting minutes database 107 may store all meeting minutes including each meeting record as a constituent element. According to another embodiment of the present invention, the meeting minutes database 107 may store not only the meeting records by subject, but also meeting records by speaker and meeting records by specific time slots generated by the automatic meeting record generation device 100 .

3 is a flowchart of a method for automatically generating a meeting record according to another embodiment of the present invention. Hereinafter, it is assumed that the automatic meeting record generation method is performed by the automatic meeting record generation device 100 . In FIG. 3 , the automatic meeting recording method of the automatic meeting recording generating device 100 will be briefly described once again, and each step will be described in detail with reference to FIGS. 4 to 7 .

Referring to FIG. 3 , the apparatus 100 for automatically generating conference records may receive voice contents of a plurality of speakers participating in the conference (S10). Also, the apparatus 100 for automatically generating a meeting record may convert the received voice content into text for each speaker and extract keywords from the converted text (S20).

Next, the apparatus 100 for automatically generating a meeting record may determine a first keyword based on the keyword extracted from the converted text (S30).

In this case, if the first keyword is determined, the automatic meeting record generating apparatus 100 may generate comment text for each speaker participating in the conference based on the determined first keyword (S40). Also, the apparatus 100 for automatically generating a conference record may generate a conference record for the first keyword, including the generated comment text for each speaker.

In step S30, the first keyword may be determined as a keyword extracted more than a threshold number of keywords extracted in step S20. For example, when keyword 1 is extracted in a predetermined amount or more from time a to end b, the apparatus 100 for automatically generating meeting records uses keywords for topics of meetings held during the period from time a to end b as keywords. 1 can be determined. Meanwhile, in order to prevent an off-topic keyword from being determined as a result of an off-topic conversation within a short period of time, the apparatus 100 for automatically generating meeting records may determine a keyword extracted over a predetermined period or a threshold number as the first keyword. there is.

Meanwhile, in the above, the method of determining the first keyword according to the number of extracted keywords has been described, but according to another embodiment of the present invention, the following various methods may be applied to the method of determining the first keyword.

One. If there is a pre-registered meeting topic

In step S20, after extracting the keyword, the automatically generating meeting record device 100 may determine whether a pre-registered meeting topic exists.

As a result of the determination, if there is a pre-registered conference topic, the automatic meeting record generation device 100 selects a keyword matching the pre-registered conference topic among the keywords extracted from the converted text in step S20 as the first keyword. can be determined by

For example, before the start of a meeting or in the middle of a meeting, a moderator or a manager of the automatic meeting record generating device 100 may register a meeting topic. The automatic meeting record creation device 100 may identify a registered meeting subject. In this case, the apparatus 100 for automatically generating a meeting record may extract keywords from the meeting subject. The apparatus 100 for automatically generating a meeting record may determine, as the first keyword, a keyword that matches the keyword extracted from the meeting topic. In this case, in order to determine whether or not to match, the automatically generating device 100 may use the above-described dictionary or search a web page.

2. meeting on the document When using keywords

In step S10, the apparatus 100 for automatically generating a meeting record may display the meeting document. In this case, in step S30, the apparatus 100 for automatically generating a meeting record may extract keywords from text on the displayed document. Also, the apparatus 100 for automatically generating a meeting record may determine, as the first keyword, a keyword that matches the keyword extracted from the text on the displayed document, among the keywords extracted from the text converted in step S20.

For example, during a meeting, the automatic meeting record generating device 100 displays a meeting document through the display device 200, and the automatic meeting record generating device 100 displays text “smart phone” on the meeting document being displayed. It is assumed that it is identified and extracted as a keyword.

In the above example, when the keyword "smart phone" of speaker A is extracted, the automatic meeting record generating device 100 matches the keyword "smart phone" of speaker A to "smart phone" extracted from the text on the conference document being displayed. can determine what is In this case, the device 100 for automatically generating a meeting record may determine “smart phone” as the first keyword.

According to another embodiment of the present invention, the apparatus 100 for automatically generating meeting records may search for a table of contents page on a displayed meeting document. The apparatus 100 for automatically generating a meeting record may identify a table of contents page, identify text on the table of contents page, and extract keywords. The Table of Contents page contains the text on the topic of the meeting, so it can increase the accuracy of keywords.

3. When using relationships between texts

In step S20, the apparatus 100 for automatically generating a meeting record may determine a relationship between the converted texts for each speaker. In this case, the automatic meeting record generation device 100 may determine relevance between the texts based on at least one of similarities between sentence structure of the converted text for each speaker and words constituting the text.

For example, the apparatus 100 for automatically generating a meeting record may recognize that speaker A asks a question and speaker B replies based on the received voice content and/or the sentence structure of the converted text. Accordingly, the apparatus 100 for automatically generating a meeting record may determine that there is a relationship between the text of speaker A and the text of speaker B.

In addition, even when a common “smart phone” is identified as a word constituting the texts of speaker A and speaker B, the automatic meeting record creation device 100 has a relationship between the text of speaker A and the text of speaker B. can also judge. In this case, even when "smart phone" in speaker A's text and "mobile phone" in speaker B's text are identified so that a semantic relationship can be recognized through a dictionary or web page search, even if they are not common words, the meeting is automatically recorded. The generating device 100 may determine that there is a relationship between the two.

As a result of determining the relevance between the texts, the automatic meeting record generation device 100 may select a text whose relevance is greater than or equal to a threshold value, and may extract a keyword from the selected text.

The apparatus 100 for automatically generating a meeting record may determine a keyword extracted from the selected text as a first keyword in step S30.

In the above embodiment, a threshold value for relevance between texts may be set in advance by an administrator of the device 100 for automatically generating meeting records. The degree of relevance may be determined by the question and the type of speaker who answers the question and/or the degree of duplication of words in which a common or semantic relationship is recognized.

For example, if there is an answer from speaker B to a question of speaker A, who is a VIP, and an answer from speaker B to a question from speaker C, the apparatus 100 for automatically generating a meeting record automatically outputs the question of speaker A and the speaker in response thereto. B's answer can be judged to have higher relevance between texts.

In another example, when word overlap between speaker A's text and speaker B's text is 3 times and word overlap between speaker A's text and speaker C's text is 1 time, the automatic meeting record generating device 100 electronically It can be judged that the relevance between the texts of is higher than the latter.

4 is an example of content for each subject according to the order of conference progress, which is referred to in some embodiments of the present invention.

Referring to FIG. 4 , a time table 401 time-sequentially shows the order of receiving audio contents according to the utterances of speakers A, B, and C during a conference. The time table 402 represents keywords X1 and X2 for the subject of the conference that change over time during the conference.

Meeting document 403 is a meeting document displayed during a meeting for keyword X1, and meeting documents 404 and 405 are meeting documents displayed during a meeting for keyword X2.

Audio contents a1, a2, and a3 are audio contents according to the speech of speaker A, audio contents b1 and b2 are audio contents according to the speech of speaker B, and audio contents c1, c2, and c3 are audio contents according to the speech of speaker C am.

The apparatus 100 for automatically generating conference records may receive voice contents of speaker A, speaker B, and speaker C like the time table 401 .

In addition, according to the order of the meeting, the automatic meeting record generating device 100 may display the meeting documents 403 , 404 , and 405 .

The apparatus 100 for automatically generating a conference record converts a1, b1, and c1, the audio contents of speakers A, B, and C, into text, extracts keywords from each text, and extracts a first keyword X1 based on the converted keywords. can At this time, the apparatus 100 for automatically generating a meeting record may identify the displayed document 403 .

Next, the automatic meeting recording device 100 converts the voice contents of speaker A, speaker B, and speaker C, a2, b2, c2, a3, and c3 into text, extracts keywords from each text, and based on this A second keyword X2 may be extracted.

In step S30, the apparatus 100 for automatically generating a meeting record may determine whether a keyword determined based on a keyword extracted from the converted text is changed from the first keyword to a second keyword.

At this time, if the keyword extracted from the text of a preset speaker among the plurality of speakers is changed by a threshold amount or more for a preset time or more, the automatic meeting record generating device 100 determines that the keyword has changed from the first keyword to the second keyword. can also judge. For example, when the keyword X1 is extracted from a preset speaker A and then the keyword X2 is extracted for more than a predetermined time, the apparatus 100 for automatically generating a meeting record may determine that the keyword has been changed. This is in consideration of the fact that changes in the keywords of the moderator or VIP affect the conference topic rather than the keywords of other speakers.

When the determined keyword is changed from the first keyword to the second keyword, the automatic meeting record generation device 100 may extract the content on the page of the displayed meeting document until the second keyword is determined. That is, the apparatus 100 for automatically generating a conference record may extract content on a page of the conference document 403 being displayed before the second keyword X2 is determined.

The apparatus 100 for automatically generating a conference record may generate a conference record for the first keyword X1 including the extracted content in step S50.

In this case, the displayed conference document may include a plurality of different documents displayed during the conference. That is, when a plurality of different documents are displayed in an arbitrary order during a meeting for the first keyword X1, the automatic meeting record generation device 100 may extract the content on the page of each displayed document, This may be reflected in the meeting record for the first keyword X1.

On the other hand, referring to the meeting document 404 and the meeting document 405 of FIG. 4 , while the meeting document 404 is displayed, the meeting document 405 is generated by reflecting the contents of the participants' writings according to the progress of the meeting. shown as an example. That is, in a state where the meeting document 404 is displayed, a participant may input writing, and in this case, the automatic meeting record generating apparatus 100 may detect that writing is started and finally completed. When a meeting is in progress with respect to the second keyword (X2), the apparatus 100 for automatically generating a meeting record may recognize a writing input performed before the second keyword is changed as a final writing input. Alternatively, when the displayed meeting document is changed from the meeting document 405 to another meeting document, the writing input performed before the change may be recognized as the final writing input.

The automatic meeting record generation device 100 may reflect writing information according to the final writing input to the meeting record.

According to another embodiment of the present invention, when a real-time conference service is being provided, the writings may be individually input through participants' terminals.

5 is an exemplary diagram for explaining a process of converting audio content of a conference participant into text and creating comment text, which is referred to in some embodiments of the present invention.

According to steps S20 and S40 of FIG. 3 , the apparatus 100 for automatically generating a meeting record may extract a keyword from voice content and generate comment text for each speaker based on the determined first keyword. Hereinafter, a method for generating comment text for each speaker in step S40 based on the first keyword using the keyword extracted in step S20 will be described.

Referring to FIG. 5 , timetables 501, 502, and 503 respectively classify received voice content by speaker. The time table 501 time-sequentially represents a1, a2, and a3, which are audio contents of speaker A. The time table 502 time-sequentially represents b1 and b2, which are the audio contents of speaker B. In addition, the time table 501 represents c1, c2, and c3, which are audio contents of speaker C, in a time-sequential manner.

In step S20, the apparatus 100 for automatically generating a meeting record may extract keyword 1, keyword 2, and keyword 3 for each audio content from the time table 501, respectively. The apparatus 100 for automatically generating a conference record may extract keywords 4 and 5 for each audio content from the time table 502 , respectively. Also, the apparatus 100 for automatically generating a meeting record may extract keyword 6, keyword 7, and keyword 8 for each audio content from the time table 503, respectively.

In operation S30 , when the first keyword is determined, the automatic meeting record generating apparatus 100 may generate comment text 511 of speaker A. The apparatus 100 for automatically generating a meeting record may generate comment text 512 of speaker B and comment text 513 of speaker C. As described above, based on the determined first keyword, the apparatus 100 for automatically generating a meeting record may generate comment text 511 by summarizing results of text conversion of the audio contents a1, a2, and a3 of speaker A. In addition, comment texts 512 and 513 may be generated as summaries of speaker B's voice contents b1 and b2 and speaker C's voice contents c1, c2, and c3, respectively.

6 is an exemplary diagram for explaining a method of identifying a keyword for a topic of an ongoing conference according to another embodiment of the present invention.

Referring to FIG. 6 , the apparatus 100 for automatically generating meeting records may determine keywords for the subject of the meeting based on the number of times per hour that keywords equal to or greater than a threshold number are extracted. In the graph 600, a case in which the first keyword (X1) is changed to the second keyword (X2) and the meeting topic is changed to the third keyword (Y) according to the order of time is illustrated as an example.

In particular, in the graph 600, the case where the second keyword (X2) and the third keyword (Y) overlap for some time is illustrated.

When such overlapping conference topics are discussed, the automatically generating conference record 100 classifies the audio content by speaker as shown in the time tables 501, 502, and 503 of FIG. 5 and then classifies the extracted keywords again, thereby Meeting records for the second keyword (X2) and the third keyword (Y) may be respectively created. In this case, the apparatus 100 for automatically generating a meeting record may generate a comment keyword for each determined keyword regardless of the speaker's speaking time and reflect the comment keyword to each conference record.

On the other hand, when there is a relationship of a threshold value or higher between the first keyword X1 and the second keyword X2, the automatically generating meeting record device 100 determines the first keyword X1 and the second keyword X2. You can also create a meeting record with a meeting record combined.

7 is an example of an automatically generated conference record, referenced in some embodiments of the present invention.

7 is an example of a meeting record 700 created according to the method of automatically generating a meeting record described above. Referring to FIG. 7 , a meeting record 700 is a meeting record for keyword X, which includes comment texts 702, 703, and 704 of speakers A, speakers B, and speaker C, and a document 705 displayed during the meeting for keyword X. , 706).

The apparatus 100 for automatically generating meeting records may store meeting records for each subject, such as the meeting record 700 . Accordingly, the user can access the meeting record 700 by inputting a keyword for a subject even if the user participates in the meeting at the end of the meeting or in the middle of the meeting.

Also, according to another embodiment of the present invention, the apparatus 100 for automatically generating a meeting record may generate a conference record based on a keyword input by a user.

In a situation where a conference is in progress, it is assumed that the automatic meeting recording device 100 converts the voice content into text after receiving the voice content in a manner similar to steps S10 and S20 of FIG. 3 described above.

In this case, the apparatus 100 for automatically generating a meeting record may receive a keyword from the user terminal. The apparatus 100 for automatically generating meeting records may extract keywords matching the received keywords from the converted text in a manner similar to step S20.

The automatic meeting record generation device 100 may generate comment text for each speaker based on the extracted keywords in a manner similar to step S40 of FIG. 3 .

Next, the apparatus 100 for automatically generating a conference record may create a conference record for the received keyword, including the generated comment text for each speaker.

Also, the automatic meeting record generation device 100 may create a meeting record including contents of a meeting document matching the received keyword. That is, the apparatus 100 for automatically generating a conference record may extract content on a page of a conference document including the keyword in response to an input keyword, and may reflect the extracted content in the conference record.

Accordingly, a user participating in a conference can access the conference record generated based on the keyword by the automatic conference record generation device 100 by inputting a keyword through the user's terminal while the conference is in progress.

In addition, by using the above keyword-based conference record generation method, a user-customized conference record can be provided in a situation where the conference topic of interest to each user may be different.

Methods according to embodiments of the present invention described so far with reference to the accompanying drawings can be performed by executing a computer program implemented as a computer readable code. The computer program may be transmitted from the first computing device to the second computing device through a network such as the Internet, installed in the second computing device, and thus used in the second computing device. The first computing device and the second computing device include both a server device, a stationary computing device such as a desktop PC, and a mobile computing device such as a notebook computer, a smart phone, and a tablet PC.

Although the embodiments of the present invention have been described with reference to the accompanying drawings, those skilled in the art can realize that the present invention can be implemented in other specific forms without changing the technical spirit or essential features. you will be able to understand Therefore, it should be understood that the embodiments described above are illustrative in all respects and not limiting.

Claims (11)

A method performed by a meeting record automatically generating device implemented as a computing device, comprising:
receiving, by the network interface of the automatic conference recording device, content of the plurality of speakers participating in the conference through terminals of the plurality of speakers;
converting the received content into text for each speaker and extracting a keyword from the converted text, by a processor of the automatic meeting record generating device;
determining, by the processor, a first keyword based on a keyword extracted from the converted text;
generating, by the processor, comment text for each speaker based on the determined first keyword; and
Generating, by the processor, a meeting record for the first keyword including the generated comment text for each speaker;
Generating a meeting record for the first keyword including the generated comment text for each speaker,
Among the first and second utterances of a first speaker among the plurality of speakers, when the first utterance includes the first keyword, a weight is applied to the first utterance to generate comment text of the first speaker. Including the steps of
The first utterance and the second utterance are temporally separated,
How to automatically create a meeting record. According to claim 1,
Determining the first keyword,
determining whether there is a pre-registered meeting topic;
When the pre-registered conference topic exists, determining a keyword matching the pre-registered conference topic as the first keyword among keywords extracted from the converted text.
How to automatically create a meeting record. According to claim 1,
Receiving the contents of the plurality of speakers,
Displaying a meeting document through a display device connected to the automatically generating meeting record device;
Determining the first keyword,
extracting keywords from text on the displayed meeting document; and
Determining, among keywords extracted from the converted text, a keyword matching a keyword extracted from text on the displayed meeting document as the first keyword,
How to automatically create a meeting record. According to claim 1,
Receiving the contents of the plurality of speakers,
Displaying a meeting document through a display device connected to the automatically generating meeting record device;
Determining the first keyword,
determining whether a keyword determined based on a keyword extracted from the converted text is changed from the first keyword to a second keyword;
When the determined keyword is changed from the first keyword to the second keyword, extracting content on a page of the conference document displayed until the second keyword is determined;
Generating a meeting record for the first keyword including the generated comment text for each speaker,
Generating a meeting record for the first keyword further comprising the extracted content,
How to automatically create a meeting record. According to claim 4,
Determining whether the keyword is changed from the first keyword to the second keyword,
Determining that the keyword is changed from the first keyword to the second keyword when a keyword extracted from the text of a preset speaker among the plurality of speakers is changed by a threshold amount or more for a preset time or more.
How to automatically create a meeting record. According to claim 4,
Generating a meeting record for the first keyword including the generated comment text for each speaker,
generating a meeting record for the first keyword by further including writing information received from terminals of the plurality of speakers through the network interface until the determined keyword is changed from the first keyword to the second keyword; including,
How to automatically create a meeting record. According to claim 1,
The step of converting the received content into text for each speaker and extracting a keyword from the converted text,
determining relevance between the converted texts for each speaker;
Selecting a text whose relevance is greater than or equal to a threshold value as a result of the determination; and
Extracting keywords from the selected text;
Determining the first keyword,
Determining a keyword extracted from the selected text as the first keyword,
How to automatically create a meeting record. According to claim 1,
Determining the first keyword,
Among the keywords extracted from the converted text, determining a keyword extracted more than a threshold amount during a preset period as the first keyword,
How to automatically create a meeting record. A method performed by a meeting record automatically generating device implemented as a computing device, comprising:
receiving, by the network interface of the automatic conference recording device, content of the plurality of speakers participating in the conference through terminals of the plurality of speakers;
converting, by a processor of the automatic meeting record generation device, the received content into text for each speaker;
receiving, by the network interface, a keyword from a user terminal;
extracting, by the processor, a first keyword matching the received keyword from among the converted text;
generating, by the processor, comment text for each speaker based on the extracted first keyword; and
generating, by the processor, a meeting record for the received keyword, including the generated comment text for each speaker;
Generating a meeting record for the first keyword including the generated comment text for each speaker,
Among the first and second utterances of a first speaker among the plurality of speakers, when the first utterance includes the first keyword, a weight is applied to the first utterance to generate comment text of the first speaker. Including the steps of
The first utterance and the second utterance are temporally separated,
How to automatically create a meeting record. According to claim 9,
Generating a meeting record for the received keyword,
generating a meeting record further comprising content of a meeting document matching the received keyword.
How to automatically create a meeting record. one or more processors;
a network interface communicating with a terminal of a speaker attending the conference;
a memory for loading a computer program executed by the processor; and
Including a storage for storing the computer program,
The computer program,
an operation for receiving contents of a plurality of speakers attending the conference;
an operation of converting the received content into text for each speaker and extracting a keyword from the converted text;
an operation of determining a first keyword based on a keyword extracted from the converted text;
an operation of generating comment text for each speaker based on the determined first keyword; and
An operation of generating a meeting record for the first keyword including the generated comment text for each speaker,
An operation of generating a meeting record for the first keyword including the generated comment text for each speaker,
Among the first and second utterances of a first speaker among the plurality of speakers, when the first utterance includes the first keyword, a weight is applied to the first utterance to generate comment text of the first speaker. It includes an operation that
The first utterance and the second utterance are temporally separated,
Device for automatic creation of meeting records.

Images