KR102476099B1 - METHOD AND APPARATUS FOR GENERATING READING DOCUMENT Of MINUTES - Google Patents

Abstract

A method and apparatus for generating a meeting minutes reading document are provided. A method for generating a meeting minutes reading document according to an embodiment of the present invention includes generating a plurality of speech unit data by dividing draft meeting minutes data for a meeting, and utterances included in each of the plurality of speech unit data according to a predetermined condition. Performing suitability evaluation on the content and selecting inappropriate speech unit data according to the result of the suitability evaluation, and reading meeting minutes using the remaining speech unit data excluding the selected inappropriate speech unit data among the plurality of speech unit data It involves generating a document.

Description

Method and device for generating meeting minutes reading document {METHOD AND APPARATUS FOR GENERATING READING DOCUMENT OF MINUTES}

The present invention relates to a method and apparatus for generating a meeting minutes reading document. More specifically, a meeting minutes reading document creation method capable of generating a meeting minutes reading document containing only remarks suitable for the agenda of the meeting by filtering out remarks included in the meeting minutes that are not suitable for the agenda of the meeting, and a method for generating the meeting minutes reading document It's about the device.

The proceedings and contents of remarks of the meetings held in the National Assembly, metropolitan local councils, or local councils are recorded by the secretary. And, it is common that the minutes recorded in this way are disclosed to the general public.

In the case of a conventional meeting minutes management method, each meeting minutes is stored as a unit. At this time, the spoken language style of the meeting minutes is managed without any change so as not to distort the intention of the speaker.

However, in the case of the conventional meeting minutes management method, since it includes all of the contents of remarks that deviate from the agenda of the meeting, contents of remarks that do not fit the gist of the question, and contents of sudden remarks that do not fit the situation, users who read the minutes There is a problem that it gives inconvenience to accurately grasp the contents of the meeting based on the remarks on the meeting agenda.

In addition, in the case of the conventional meeting minutes management method, it only provides information on who spoke about a specific remark or what type of remark the specific remark is, and intuitively determines whether the remark is related to the agenda. Since it is difficult to grasp, there is a problem that users have to read and check the contents of the remarks to the end and make a judgment.

Korean Patent Publication No. 2014-0077514

The technical problem to be solved by the present invention is to create a plurality of speech unit data by dividing the draft meeting minutes data for the meeting, and select inappropriate speech unit data through suitability evaluation for each of the plurality of speech data, It is to provide a method and apparatus for generating a meeting minutes reading document capable of filtering out unsuitable remarks.

Another technical problem to be solved by the present invention is a method for generating a meeting minutes reading document that allows users to intuitively check whether it is appropriate for a subject by displaying the result of suitability evaluation for each remark in a meeting minutes reading document, and a method for generating the meeting minutes reading document. to provide the device.

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, a method for generating a meeting minutes reading document according to an embodiment of the present invention is a method performed by a computing device, which divides draft meeting minutes data for a meeting to generate a plurality of speech unit data. The step of performing suitability evaluation on the utterances included in each of the plurality of speech unit data according to a predetermined condition, and selecting inappropriate speech unit data according to the suitability evaluation result and among the plurality of speech unit data. and generating a meeting minutes reading document using the remaining speech unit data excluding the selected inappropriate speech unit data.

In some embodiments, the selecting of inappropriate speech unit data according to the suitability evaluation may include extracting a keyword from each of the plurality of speech unit data and comparing the extracted keywords to determine a similarity between the keywords. and selecting speech unit data having the determined similarity equal to or less than a preset reference value as inappropriate speech unit data.

In some embodiments, the plurality of speech unit data includes first speech unit data including the speech content of the first speaker and second speech unit data including the speech content of the second speaker, and the suitability evaluation The step of selecting inappropriate speech unit data according to the result may include, when the speech content included in the first speech unit data includes information on the second speaker and a keyword for selecting the inappropriate speech unit data, the first speech unit data includes the second speaker data. 2 Selecting speech unit data as the inappropriate speech unit data.

In some embodiments, the selecting of inappropriate speech unit data according to the suitability evaluation result may include determining whether the microphones of the first speaker and the second speaker are activated using log data on microphones used for speaking in the conference. When it is determined that the microphone of the second speaker is activated while the microphone of the first speaker is activated, the speech unit data including the contents of speech input through the microphone of the second speaker A step of selecting inappropriate speech unit data may be included.

In some embodiments, the method of generating a meeting minutes reading document further includes obtaining image data of a plurality of speakers attending the meeting from a photographing device, and selecting inappropriate speech unit data according to a result of the suitability evaluation. In the step, by using the video data, at the time when the specific speaker speaks, the amount of expression change and the behavior pattern of other speakers are analyzed, and the contents of the speech of the specific speaker are analyzed using the analysis result of the amount of change in expression and the behavior pattern. A step of evaluating suitability for included speech unit data may be included.

In some embodiments, the selecting of inappropriate speech unit data according to a result of the suitability evaluation may include determining whether or not a plurality of speakers attending the conference are speaking targets by using progress scenario information for the conference; and The method may include selecting, as the inappropriate speech unit data, speech unit data including speech content of a speaker determined not to be the speech target from among the plurality of speech unit data.

In some embodiments, the generating of the meeting minutes reading document includes only the first meeting minutes reading document including all of the plurality of speech unit data and the speech unit data other than the inappropriate speech unit data among the plurality of speech unit data. Generating a second meeting minutes reading document including, and when receiving a reading request for the meeting minutes reading document, one of the first meeting minutes reading document or the second meeting minutes reading document according to the reading request Optionally providing may be included.

In some embodiments, the generating of the meeting minutes reading document may include, when a request for reading the meeting minutes reading document is received, the contents of the plurality of speech unit data and the suitability evaluation result for each of the plurality of speech unit data. It may include a step of outputting together.

In order to solve the above technical problem, an apparatus for generating a meeting minutes reading document according to another embodiment of the present invention includes a communication interface for receiving draft meeting minutes data for a meeting, a storage for storing software for generating and providing a meeting minutes reading document, A memory into which a plurality of instructions are loaded according to execution of the software and a processor that executes the plurality of instructions, wherein the plurality of instructions divide the draft meeting minutes data to generate a plurality of speech unit data an instruction for performing suitability evaluation on the utterances included in each of the plurality of speech unit data according to a predetermined condition, and selecting inappropriate speech unit data according to the suitability evaluation result, and among the plurality of speech unit data An instruction for generating a meeting minutes reading document using the remaining speech unit data excluding the selected inappropriate speech unit data may be included.

1 is a diagram illustrating a system for generating and providing a meeting minutes reading document according to an embodiment of the present invention.
2 is a flowchart illustrating a method of generating a meeting minutes reading document according to another embodiment of the present invention.
3 is a diagram exemplarily illustrating a form in which a plurality of speech unit data is generated from draft meeting minutes data according to some embodiments.
4 is a flowchart illustrating a method of selecting inappropriate speech unit data through comparison of similarities between keywords, in an embodiment.
FIG. 5 is a diagram illustrating a process of selecting inappropriate speech unit data according to the method of selecting inappropriate speech unit data through comparison of similarity between keywords described with reference to FIG. 4 .
6 is a flowchart illustrating a method of selecting inappropriate speech unit data using microphone activation information in another embodiment.
FIG. 7 is a diagram illustrating a process of selecting inappropriate speech unit data according to the method of selecting inappropriate speech unit data using the microphone activation information described with reference to FIG. 6 .
8 is a flowchart illustrating a method of selecting inappropriate speech unit data according to whether a keyword exists or not, according to another embodiment.
FIG. 9 is a diagram illustrating a process of selecting inappropriate speech unit data according to the method of selecting inappropriate speech unit data according to the presence or absence of a keyword described with reference to FIG. 8 .
10 is a flowchart illustrating a method of selecting inappropriate speech unit data according to a speaker's facial expression and behavior in another embodiment.
FIG. 11 is a diagram illustrating a process of selecting inappropriate speech unit data according to the method of selecting inappropriate speech unit data according to a speaker's expression and behavior described with reference to FIG. 10 .
12 to 14 are diagrams illustratively illustrating a form in which a meeting minutes reading document generated through the method for generating a meeting minutes reading document described with reference to FIG. 2 is output to a meeting minutes reading device.
15 is a hardware configuration diagram of a meeting minutes reading document generating device according to another embodiment 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.

Hereinafter, some embodiments of the present invention will be described with reference to the drawings.

1 is a diagram illustrating a system for generating and providing a meeting minutes reading document according to an embodiment of the present invention.

Referring to FIG. 1, the system for generating and providing a meeting minutes reading document according to an embodiment of the present invention includes a meeting minutes reading document generating device 100, a meeting recording device 200, a log management server 300, a plurality of microphones 400 -1 to 400-n) and the meeting minutes reading device 500.

Here, the meeting minutes reading document generation and provision system shown in FIG. 1 is according to an embodiment, and its components are not limited to the embodiment shown in FIG. 1, and may be added, changed, or deleted as necessary. Note the

In addition, the meeting minutes reading document generation and provision system shown in FIG. 1 shows functional elements in which each component is functionally divided, and any one or more components may be implemented by integrating each other in an actual physical environment,

First, the meeting recording device 200 may generate data related to meeting minutes for a meeting and transmit the data to the meeting minutes reading document generating device 100 . More specifically, the meeting recording device 200 may generate draft meeting minutes data and provide the meeting minutes reading document generating device 100 .

Here, a council is a procedure in which two or more participants exchange opinions and information on one or more proposed agendas and proceed with decision-making. Meetings according to some embodiments of the present invention may include, but are not limited to, plenary sessions, legislation and judicial committees, relevant committees, special committees, standing committees, administrative affairs audits and investigations, and governmental audits and investigations. In addition, participants in a meeting according to an embodiment of the present invention may include a chairperson or a chairperson, a member of the council or a member, but are not limited thereto.

In addition, here, the draft meeting minutes is a document in which the progress of the meeting, the contents of the meeting participants' remarks, or the meeting results are recorded, and the draft meeting minutes data refers to string data, audio data, and video data including such information. However, it is not limited thereto.

Next, the log management server 300 may generate and transmit log data generated according to the progress of the meeting to the meeting minutes management server 300 .

Here, the log data is character string data in which a log of an event generated in the course of a conference is recorded. For example, the log data may be character string data recorded with information about the time when the microphones 400-1, 400-2, and 400-n of the speakers are activated, but is not limited thereto.

The meeting minutes reading document generating device 100 may generate a plurality of speech unit data by dividing the draft meeting minutes data provided from the meeting recording device 200 using the log data provided from the log management server 300 .

In addition, the meeting minutes reading document generating device 100 may perform a suitability evaluation to determine whether each of the plurality of speech unit data is compatible with the agenda of the meeting, and through this process, by selecting inappropriate speech unit data, You can filter speech unit data about speeches that do not fit the agenda.

Next, the meeting minutes reading device 500 may receive and output the meeting minutes reading document generated from the meeting minutes reading document generating device 100 . For example, the meeting minutes reading device 500 according to an embodiment of the present invention includes a desktop, a workstation, a server, a laptop, a tablet, and a smart phone. ) or a phablet, but is not limited thereto, and is not limited thereto, such as a portable multimedia player (PMP), a personal digital assistant (PDA), or an e-book reader (E-Book Reader). ) and the like.

In one embodiment, the meeting minutes reading document is a document generated by the meeting minutes reading document generating device 100 at the request of the meeting minutes reading device 400, and includes the progress of the meeting, the contents of the speech, the time of the meeting, and the meeting result. can

In addition, the meeting minutes reading document may be a web document in a format such as HTML (Hyper Text Markup Language) or XML (eXtensible Markup Language), but is not limited thereto, and may be a PDF (Portable Document Format) document. Hereinafter, with reference to FIGS. 2 to 13 , a method for generating a meeting minutes reading document generating device 100 will be described in detail.

2 is a flowchart illustrating a method of generating a meeting minutes reading document according to another embodiment of the present invention. A method for generating a meeting minutes reading document according to another embodiment of the present invention may be performed by a computing device. For example, the computing device may be the meeting minutes reading document generating device 100 of FIG. 1 . Hereinafter, if the description of the subject of the operation is omitted in the description related to this embodiment, it will be understood that the subject is the meeting minutes reading document generating device 100 of FIG. 1 .

Referring to FIG. 2 , first, draft meeting minutes data is obtained (S110). For example, the meeting minutes reading document generating device 100 may receive text, voice, or video-based draft meeting minutes data generated from the meeting recording device 200 through a network. However, it is not limited thereto.

Next, by dividing the draft meeting minutes data, a plurality of speech unit data is generated (S120). More specifically, the meeting minutes reading document generating device 100 may generate a plurality of speech unit data by dividing the draft meeting minutes data according to a predetermined criterion. For example, the meeting minutes reading document generating device 100 may identify a speaker for the remarks by using the microphone activation time of each of a plurality of speakers included in the log data, and generate draft meeting minutes data with the identified speaker as a mood. It is possible to generate a plurality of speech unit data by dividing according to the contents of the speech of the speaker. However, it is not limited to this, and the apparatus 100 for generating a meeting minutes reading document divides the draft meeting minutes data according to the speaker's remarks, such as a division method according to a page progress signal from a prompter and a division method according to a user input signal, and speaks. Any method capable of generating unit data can be applied.

As shown in FIG. 3, in the case of the conventional meeting minutes, the speaker's position (1), the speaker's name (2), a separator (4) for distinguishing the speaker's name (2) and the contents of the remarks (3), and the contents of the remarks (3) is described successively. In addition, the conventional speaker identification method distinguishes the speaker's name (2) from the speech content (3) based on the separator (4) included in the conventional meeting minutes data.

However, the conventional speaker identification method cannot correctly identify the speaker's name (2) and the statement (3) when the speaker's position (1) or speaker's name (2) includes the separator (4). . For example, in the conventional speaker identification method, when the delimiter (4) is a blank character, the speaker's position (1) is "local transportation director" and the speaker's name (2) is "XXX", but the speaker's It is identified that the position (1) is "province", the speaker's name (2) is "Director of Transportation", and the content of remarks (3) is "XXX first, assuming related laws...".

However, the meeting minutes reading document apparatus 100 according to an embodiment of the present invention uses speaker information (5 ) and the speech content (3), there is an advantage in that the problems of the conventional speaker identification method can be solved.

Next, suitability evaluation is performed on the speech content included in each of the plurality of speech unit data generated in step S120 (S130), and inappropriate speech unit data is selected from among the plurality of speech unit data according to the suitability evaluation result (S130). S140).

Here, suitability evaluation means evaluating and quantifying how appropriate the speech of a specific speaker is to the agenda of the meeting through the suitability evaluation algorithm. Hereinafter, with reference to FIGS. 3 to 11 , various embodiments of a method for selecting inappropriate speech unit data through the suitability evaluation by the meeting minutes reading document generating apparatus 100 will be described.

4 is a flowchart illustrating a method of selecting inappropriate speech unit data through comparison of similarities between keywords, in an embodiment.

Referring to FIG. 4 , in one embodiment, the apparatus 100 for generating a meeting minutes reading document may select inappropriate speech unit data by comparing keywords extracted from each of a plurality of speech unit data.

First, a keyword is extracted from each of the plurality of speech unit data generated in step S120 (S210).

Here, the keyword is a main word included in the speech content. For example, the keywords may include, but are not limited to, keywords, core words, frequently occurring words, conference names, sessions, sessions, agenda names, and the like.

Next, similarities between keywords extracted from each of a plurality of speech unit data in step S210 are calculated (S220). For example, as shown in FIG. 5, the apparatus 100 for generating a meeting minutes reading document is a keyword of "A, B, C, D, E" from the speech unit data 20b including the contents of the second speaker's remarks ( 30b) can be extracted, and the keyword 30c of “A, F, G, H I” can be extracted from the speech unit data 20c including the speech contents of the third speaker, and the speech contents of the fourth speaker Keywords 30d of “B, C, D, E, P” may be extracted from the included speech unit data 20d. Thereafter, the meeting minutes reading document generating device 100 sets the keyword set “A, B, C, D, E, G, K, L, M” and the first speaker and the second speaker according to the agenda of the meeting, that is, the subject. And the keywords extracted from the speech unit data 20b, 20c, and 20d including the speech contents of the third speaker may be compared, and the similarity determination result 50 may be calculated through the comparison result. The similarity judgment result 50 shown in 5 is shown in the form of a percentage (%), but is not limited thereto.

Keyword extraction method Regarding the method of judging the similarity between keywords, "Improved automatic keyword extraction given more linguistic knowledge", Anette Hulth, 2003, "Study on single document keyword extraction using term clustering", Han Seung-hee, 2010 and "Clustering technique Expansion of User Queries According to Keyword Similarity Ranking Algorithm Using "," Sang-Hoon Lee et al., 2003, etc. may be referenced. However, it is preferable to use an algorithm capable of obtaining a digitized similarity judgment result in order to compare the similarity judgment result with a reference value.

Next, the similarity between keywords calculated in step S220 is compared with a preset reference value (S230).

In step S230, speech unit data whose similarity between keywords is less than or equal to a reference value is selected, speech unit data whose similarity is less than or equal to the reference value is selected as unsuitable speech unit data (S240), and speech unit data whose similarity exceeds the reference value is suitable. It is selected as speech unit data (S250). For example, when the preset reference value is 70%, the apparatus 100 for generating a meeting minutes reading document generates speech unit data including the contents of a speech of a third speaker whose similarity is 70% or less according to the similarity determination result calculated in step S220. (20c) can be selected as inappropriate speech unit data. In addition, the meeting minutes reading document generating device 100 converts speech unit data 20b including the speech content of the first speaker having a similarity of 70% or more and speech unit data 20d including the speech content of the fourth speaker to appropriate utterances. Can be selected as unit data. However, it is not limited thereto, and the apparatus 100 for generating a meeting minutes reading document may select only inappropriate speech unit data through similarity comparison, and may select all of the remaining speech unit data that are not selected as suitable speech unit data.

Through this, there is an advantage in that it is possible to shorten the time required for suitability evaluation performed on each of a plurality of speech unit data by reducing an unnecessary process (a process of selecting suitable speech unit data). Next, a method of selecting inappropriate speech unit data according to another embodiment will be described with reference to FIG. 6 .

6 is a flowchart illustrating a method of selecting inappropriate speech unit data using microphone activation information in another embodiment.

Referring to FIG. 6 , in another embodiment, the apparatus 100 for generating a meeting minutes reading document may select inappropriate speech unit data using microphone activation information.

First, after a plurality of speech unit data is generated through step S120, a check operation is started to check whether the microphone of the first speaker is activated (S310).

Next, according to the start of the microphone activation confirmation operation in step S310, it is checked whether the microphone of the first speaker is activated using the microphone activation time information included in the log data (S320), and the first speaker's microphone is activated through step S320. When it is determined that the microphone is activated, a confirmation operation is initiated to determine whether the second speaker's microphone is activated (S330).

Next, in response to the start of the microphone activation confirmation operation in step S330, it is checked whether the second speaker's microphone is activated using the microphone activation time information included in the log data (S340), and the second speaker's microphone is activated through step S340. When it is confirmed that the microphone is activated, the second speech unit data including the speech content input through the second speaker's microphone is selected as inappropriate speech unit data (S350).

That is, as shown in FIG. 7, the meeting minutes reading document generating device 100 is in a state in which the microphone of the first speaker is activated, that is, in a state in which the first speaker speaks on the agenda of the meeting, the second speaker When the microphone of is activated and speech contents are input, the second speech unit data 20b including the speech contents of the second speaker may be selected as inappropriate speech unit data.

On the other hand, if it is confirmed through step S340 that the second speaker's microphone is not activated, the first speech unit data including the speech input through the first speaker's microphone is selected as suitable speech unit data (S360). Next, a method of selecting inappropriate speech unit data according to another embodiment will be described with reference to FIG. 8 .

8 is a flowchart illustrating a method of selecting inappropriate speech unit data according to whether a keyword exists or not, according to another embodiment.

Referring to FIG. 8 , in another embodiment, the apparatus 100 for generating a meeting minutes reading document, when the first speaker's remarks include the second speaker's name and a keyword for selecting inappropriate speech unit data, the second speaker Speech unit data including the contents of speech may be selected as inappropriate speech unit data.

First, after a plurality of speech unit data is generated through step S120, a keyword is extracted from the first speech unit data of the first speaker (S410). Here, as the keyword extraction method, the same method as the method of extracting the keyword in the keyword extraction step (S210) shown in FIG. 4 may be applied.

Next, it is determined whether the keywords for selecting second speaker information and inappropriate speech unit data are included among the keywords extracted in step S410 (S420). For example, as shown in FIG. 9 , among the information of the second speaker in the first speech unit data 20c including the contents of the first speaker, the second speaker's name and "unnecessary", "refrain", " When the keyword 30c for selecting unsuitable speech unit data, such as "refrain", is included, the meeting minutes reading document generating device 100 converts the second speech unit data 20b including the contents of the second speaker into unsuitable speech. It can be selected as speech unit data. Here, only "unnecessary", "restraint", and "refrain" are described as keywords for selecting inappropriate speech unit data, but are not limited thereto, and are not limited thereto. Accordingly, various keywords may be registered in advance. Next, a method for selecting inappropriate speech unit data according to another embodiment will be described with reference to FIG. 10 .

10 is a flowchart illustrating a method of selecting inappropriate speech unit data according to a speaker's facial expression and behavior in another embodiment.

Referring to FIG. 10 , in another embodiment, an apparatus 100 for generating a meeting minutes reading document analyzes facial expression changes and behavior patterns of a plurality of speakers participating in a meeting, and according to the analysis result, utterances including the current speaker's remarks. The suitability for unit data can be evaluated.

First, after a plurality of speech unit data is generated through step S120, image data for a plurality of speakers is obtained (S510). For example, as shown in FIG. 11, the meeting minutes reading document generating device 100 is connected to a separately provided photographing device 50 through a network and photographs the faces and actions of a plurality of speakers through the photographing device 50. image data can be obtained. However, it is not limited thereto. In addition, although one photographing device 50 is shown in FIG. 11 as a form of photographing a plurality of speakers, it is not limited thereto, and in order to perform more accurate image data analysis, a photographing device 50 is provided for each speaker. may be provided.

Next, by analyzing the image data for each speaker captured in step S510, the amount of expression change and behavior pattern for each speaker are analyzed (S520).

Here, the method for analyzing the amount of facial expression change and behavior pattern is "Recognizing action units for facial expression analysis", Y.-I. Various literatures such as Tian et al. 2, 2001 and "Design and Construction of a 3D Gesture Database for Human Behavior Analysis", Noh Myeong-cheol et al. 5, 2005 may be referred to. However, in order to compare the change in facial expression with a reference value, it is preferable to use an algorithm in which the amount of change in facial expression can be obtained as a numerical value.

Next, it is determined whether or not abnormality in the expression change amount and behavior pattern of the speaker has occurred according to the analysis result of the facial expression change amount and behavior pattern analyzed in step S520 (S530). For example, as shown in FIG. 11 , the meeting minutes reading document generating device 100 generates a second speaker 40-2 and a third speaker 40-3 at the time when the first speaker 40-1 speaks. ), and the amount of change in expression and behavior pattern of each speaker may be analyzed by analyzing the image data of the fourth speaker 40-4 and the fifth speaker 40-5. At this time, the facial expression changes of the second speaker 40-2, the third speaker 40-3, the fourth speaker 40-4, and the fifth speaker 40-5 are 7, 5, 9, and 6, respectively. , When the reference value is 5, the meeting minutes reading document generating device 100 includes the second speaker 40-2, the third speaker 40-3, the fourth speaker 40-4, and the fifth speaker 40-2. It can be determined that an abnormality has occurred by determining that all facial expression changes in 5) exceed the reference value. In addition, the meeting minutes reading document generating device 100 determines the behavior patterns of the second speaker 40-2, the third speaker 40-3, the fourth speaker 40-4, and the fifth speaker 40-5. Analyze, compare the behavioral pattern with the preset behavioral pattern, and analyze the behavior of the second speaker 40-2, the third speaker 40-3, the fourth speaker 40-4, and the fifth speaker 40-5. When the behavioral pattern matches a preset behavioral pattern, it may be determined that an abnormality has occurred.

Here, the preset behavioral pattern means a behavioral pattern that speakers may show when a statement does not fit the agenda of the meeting, and may be information registered in advance before analyzing the image data.

Next, if it is determined through step S530 that an abnormality has occurred in the facial expression change and behavioral pattern of the speaker, speech unit data including the contents of the speech at the time when the abnormality occurred, that is, the contents of the first speaker The first speech unit data is selected as inappropriate speech unit data (S540).

On the other hand, if it is determined that there is no abnormality in the facial expression change and behavior pattern of the speaker through step S530, the first speech unit data including the contents of the first speaker's speech is selected as suitable speech unit data (S550).

Returning to FIG. 2 again, when inappropriate speech unit data is selected from among a plurality of speech unit data in step S140, a meeting minutes reading document is created using the remaining speech unit data excluding the inappropriate speech unit data selected in step S140 ( S150).

In one embodiment, the meeting minutes reading document generating device 100 includes a first meeting minutes reading document including all of a plurality of speech unit data and a second speech unit data including only the remaining speech unit data excluding inappropriate speech unit data from among the plurality of speech unit data. It is possible to generate a meeting minutes reading document, and when receiving a request for reading the meeting minutes reading document through the meeting minutes reading device 500, one of the first meeting minutes reading document or the second meeting minutes reading document according to the reading request is a meeting minutes reading document. can optionally be provided. For example, when a request for reading the first meeting minutes reading document 60a is received from the meeting minutes reading device 100, the meeting minutes reading document generating device 100 has a plurality of speech unit data 20a as shown in FIG. , 20b, 20c) can provide a first meeting minutes reading document (60a). In addition, when a request for reading the second meeting minutes reading document 60b is received from the meeting minutes reading device 100, the meeting minutes reading document generating device 100, as shown in FIG. 13, a plurality of speech unit data 20a, A second meeting minutes reading document 60b including only the remaining speech unit data 20a and 20b excluding the inappropriate speech unit data 20c of 20b and 20c may be provided.

In one embodiment, the meeting minutes reading document generating device 100 includes a plurality of speech unit data 20a, 20b, and 20c including speech contents and information 40 of the speaker who uttered the speech contents included in each speech unit data. ) can be matched and output together. However, it is not limited thereto.

Through this, there is an advantage in that a user browsing the minutes reading document can not only check the contents of the speaker's remarks, but also easily obtain information about the speaker of the corresponding remarks.

In one embodiment, the meeting minutes reading document generating device 100 outputs all of the plurality of speech unit data as shown in FIG. 14, but displays the speech unit data selected as inappropriate speech unit data in the form of a hidden display box 63. A third meeting minutes reading document to be output may be provided, and when a user input is input to the hidden display box 63 through the meeting minutes reading device 500, inappropriate speech unit data may be output as shown in FIG. 13 have.

Through this, there is an advantage in that the user can selectively receive inappropriate speech unit data through a simple input button operation without sending a separate reading request to provide a different type of meeting minutes reading document.

In one embodiment, when the meeting minutes reading document generating device 100 receives a reading request for the meeting minutes reading documents 60a and 60b as shown in FIGS. 12 and 13, a plurality of speech unit data 20a and 20b , 20c) and the suitability evaluation result 62 for each of the plurality of speech unit data 20a, 20b, and 20c may be output together.

Through this, there is an advantage in that it is possible for a user reading a meeting minutes reading document to intuitively determine how suitable the contents of the remarks are to the agenda of the meeting.

The methods according to the embodiments of the present invention described so far 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 physical server belonging to a server pool for a cloud service, and a fixed computing device such as a desktop PC.

The computer program may be stored in a recording medium such as a DVD-ROM or a flash memory device. Referring to FIG. 15, the hardware configuration of the apparatus 100 for generating a meeting minutes reading document according to another embodiment of the present invention will be described.

15 is a hardware configuration diagram of a meeting minutes reading document generating device according to another embodiment of the present invention.

Referring to FIG. 15, the apparatus 100 for generating a meeting minutes reading document according to another embodiment of the present invention includes one or more processors 110, a communication interface 120, and a computer program 141 executed by the processor 110 It may include a memory 130 for loading, a storage 140 for storing computer programs and various data, and a bus 150. Here, in FIG. 15, only components related to the embodiment of the present invention are shown. Therefore, those skilled in the art to which the present invention pertains can know that other general-purpose components may be further included in addition to the components shown in FIG. 15 .

The processor 110 controls the overall operation of each component of the meeting minutes reading document generating device 100 . The processor 110 includes a Central Processing Unit (CPU), a Micro Processor Unit (MPU), a Micro Controller Unit (MCU), a Graphic Processing Unit (GPU), or any type of processor well known in the art of the present invention. It can be. Also, the processor 110 may perform an operation for at least one application or program for executing a method according to embodiments of the present invention. The meeting minutes reading document generating device 100 may include one or more processors.

The communication interface 120 supports wired and wireless Internet communication of the meeting minutes reading document generating device 100 . Also, the communication interface 120 may support various communication methods other than Internet communication. To this end, the communication interface 120 may include a communication module well known in the art. In some embodiments, communication interface 120 may be omitted.

Memory 130 stores various data, commands and/or information. Memory 130 may load computer program 141 from storage 140 to execute methods/operations according to various embodiments of the present invention. When the computer program 141 is loaded into the memory 130, the processor 110 may perform the method/operation by executing one or more instructions constituting the computer program 141. The memory 130 may be implemented as a volatile memory such as RAM, but the technical scope of the present invention is not limited thereto.

The storage 140 may non-temporarily store the computer program 141 . When the meeting minutes reading document generation process is performed through the meeting minutes reading document generating device 100, the storage 140 may store various data necessary for the meeting minutes reading document generating process. For example, log data and image data may be stored in order to evaluate the suitability of draft meeting minutes data, a plurality of speech unit data divided from draft meeting minutes data, and the speaker's remarks. To this end, the speech unit database 142 and A video conference transcript database 143 may be included.

The storage 140 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.

Computer program 141 may include one or more instructions that when loaded into memory 130 cause processor 110 to perform methods/operations in accordance with various embodiments of the invention. That is, the processor 110 may perform the method/operation according to various embodiments of the present disclosure by executing the one or more instructions.

In one embodiment, the computer program 141 divides the draft meeting minutes data to generate a plurality of speech unit data, performs suitability evaluation on the speech content included in each of the plurality of speech unit data according to a predetermined condition, and , It may include an instruction for selecting inappropriate speech unit data according to the result of the conformance evaluation and an instruction for generating a meeting minutes reading document using the remaining speech unit data except for the inappropriate speech unit data selected from among a plurality of speech unit data.

The bus 150 provides a communication function between components of the meeting minutes reading document generating device 100 . The bus 150 may be implemented in various types of buses such as an address bus, a data bus, and a control bus.

So far, various embodiments of the present invention and effects according to the embodiments have been described with reference to FIGS. 1 to 15 . Effects according to the technical idea of the present invention are not limited to the effects mentioned above, and other effects not mentioned will be clearly understood by those skilled in the art from the description below.

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

Claims (6)

In a method performed by a computing device,
A plurality of speech data is generated by dividing the draft meeting minutes data for the meeting, and the plurality of speech unit data includes first speech unit data including the speech contents of the first speaker and second speech unit data including the speech contents of the second speaker. including speech unit data;
When the content of speech included in the first speech unit data includes both the information on the second speaker and a keyword for selecting inappropriate speech unit data, the second speech unit data is selected as the inappropriate speech unit data; ,
selecting the second speech unit data as suitable speech unit data when at least one of information on the second speaker and the keyword is not included in the speech content included in the first speech unit data; and
Excluding the inappropriate speech unit data from among the plurality of speech unit data, generating a meeting minutes reading document using the suitable speech unit data,
How to create a meeting minutes review document. According to claim 1,
The step of generating the meeting minutes reading document,
Generating a first meeting minutes reading document including all of the plurality of speech unit data and a second meeting minutes reading document including only the remaining speech unit data excluding the inappropriate speech unit data among the plurality of speech unit data,
How to create a meeting minutes review document. According to claim 2,
The step of generating the meeting minutes reading document,
When receiving a request for reading the meeting minutes reading document, selectively providing one of the first meeting minutes reading document or the second meeting minutes reading document according to the reading request,
How to create a meeting minutes review document. According to claim 1,
The step of generating the meeting minutes reading document,
When receiving a request for reading the meeting minutes reading document, outputting together the contents of the plurality of speech unit data and the suitability evaluation result for each of the plurality of speech unit data,
How to create a meeting minutes review document. In a method performed by a computing device,
generating a plurality of speech unit data by dividing the draft meeting minutes data for the meeting;
determining whether the microphones of the first speaker and the second speaker are activated by using log data of microphones used for speaking in the conference;
When it is determined that the microphone of the second speaker is activated during the time period during which the microphone of the first speaker is activated, speech unit data including speech content input through the microphone of the second speaker is converted into inappropriate speech unit data. select as,
When it is determined that the microphone of the second speaker is not activated during the time period during which the microphone of the first speaker is activated, speech unit data including the utterance input through the microphone of the first speaker is converted to an appropriate speech unit. Selecting as data; and
Excluding the inappropriate speech unit data from among the plurality of speech unit data, generating a meeting minutes reading document using the suitable speech unit data,
How to create a meeting minutes review document. In a method performed by a computing device,
generating a plurality of speech unit data by dividing the draft meeting minutes data for the meeting;
obtaining image data of a plurality of speakers attending the conference from a photographing device;
Using the video data, the amount of change in facial expressions and behavioral patterns of other speakers are analyzed at the time when a specific speaker speaks, and if an abnormality occurs in the amount of change in facial expressions and behavioral patterns of other speakers according to the analysis result, the specific speaker's speech Selecting speech unit data including content as inappropriate speech unit data; and
Generating a meeting minutes reading document using the remaining speech unit data excluding the selected inappropriate speech unit data among the plurality of speech unit data,
How to create a meeting minutes review document.

Images