KR20220166465A - Meeting minutes creating system and method using multi-channel receiver - Google Patents

Abstract

A meeting minutes creating system using a multi-channel receiver according to one embodiment of the present invention comprises: at least one microphone installed individually for each meeting participant to receive the participant's voice; and a multi-channel receiver connected to the microphone to receive a voice signal for the voice input from the microphone through an individual channel for each microphone, and connected to a user terminal to transmit the voice signal, wherein the user terminal is connected to the multi-channel receiver, a meeting minutes management program is installed and executed in the user terminal to create an individually recorded voice file for each received voice signal, and the meeting minutes management program is a program for creating a meeting minutes file by performing STT conversion on the voice files and modifying and managing contents of meeting minutes. The meeting minutes creating system can create voice files for voice signals, extract text for voice to create meeting minutes, and provide management functions for the meeting minutes.

Description

Meeting minutes creating system and method using multi-channel receiver

The present invention relates to a system and method for generating meeting minutes, and more particularly, to a system and method for generating meeting minutes by individually collecting audio signals from a plurality of microphones and performing STT conversion on the recorded voice to generate meeting minutes. .

There is a meeting minutes system for automatically recording voice conferences for active communication between members due to various conditions such as non-face-to-face environments and situations in which meeting contents need to be recorded.

In addition, in the meeting minutes system, it is necessary to store the records of the ongoing meeting in the database, and when the meeting period ends, the meeting moderator can display the meeting result data on the mobile server and web server meeting result window and store it in the database at the same time. .

In the case of the existing meeting minutes system service, the user separately uploads the recorded meeting contents (voice file) to the server and then receives the converted text file, and a network connection is essential for this.

In addition, a function of converting and recording conversation content into text in real time is provided, and after the recording is completed, a separate post-processing process is performed to provide a conversation record separated by the speaker.

In addition, since a separate microphone is not configured, the quality of the converted text result is different depending on various environmental factors such as the type of microphone used by the user to record conversations and the distance between the user and the microphone.

Furthermore, since one voice data is created by merging the voices input by the microphone, and the speaker is distinguished through the voice separation process and the text for the voice is extracted, the voice recognition rate is significantly lowered during the voice separation process, and when the speakers speak at the same time There was a problem in that it was difficult to create proper meeting minutes because only loud sounds were separated and small sounds were misrecognized as noise, so accurate voice recognition and text extraction by separation were not performed for each speaker.

Therefore, in order to solve the above-described problem, it is possible to create a voice file for a voice signal by dividing it into an independent microphone, extract text for the voice by dividing it by speaker, and create meeting minutes, and various management functions such as editing the meeting minutes Research on a meeting minutes generation system and method using a multi-channel receiver that can provide a need has become necessary.

Korean Registered Patent No. 10-1995443 (registered on June 26, 2019)

An object of the present invention is to create a voice file for a voice signal by dividing it into an independent microphone, extract text for the voice by dividing it by speaker, generate meeting minutes, and provide various management functions such as editing the meeting minutes. It is to provide a system and method for generating meeting minutes using a multi-channel receiver.

A meeting minutes generation system using a multi-channel receiver according to an embodiment of the present invention includes at least one microphone installed individually for each conference participant to receive the participant's voice; A multi-channel receiver connected to the microphone to receive a voice signal for voice input from the microphone as an individual channel for each microphone and to transmit a voice signal, wherein the user terminal includes the multi-channel receiver A meeting minutes management program that is connected to and creates individually recorded voice files for each received voice signal is installed and executed. The meeting minutes management program creates a meeting minutes file by performing STT conversion on the voice file, Characterized in that it is a program for modifying and managing.

In the above, the meeting minutes management program includes a collecting unit that collects a voice signal corresponding to the voice signal from the multi-channel receiver and transfers the collected voice to the STT conversion unit; an STT conversion unit that includes a speech recognition model and performs text conversion on speech; a generation unit for generating a meeting minutes file by collecting the text extracted by the conversion from the STT conversion unit; and a management unit that reads the meeting minutes file generated by the generating unit and provides a function of modifying or deleting the meeting minutes upon a user's request.

In the above, a server interlocking unit for initially downloading a meeting minutes management program from a management server communicating with a user terminal through a wired/wireless communication network or for storing and managing generated meeting minutes based on a web is further included.

In the above, the management server includes a communication unit capable of communicating with a user terminal in which the meeting minutes management program is installed to receive the meeting minutes through wired or wireless communication; In order to store and manage the meeting minutes collected through the communication unit in the database, member information is provided and stored to operate as a membership system, log-in authentication for the meeting minutes management program and database storage function is performed based on the member information, and the meeting minutes management A meeting minutes management unit that supports a function of generating or editing meeting minutes by communicating based on a web program on behalf of the program; a statistical learning unit that classifies the meeting minutes by meeting, place, and time, stores and manages the minutes in a database, and generates and provides statistical information on various meeting minutes to users; It is installed and executed in the user terminal and further includes an app interlocking unit providing an interface compatible with a server interlocking unit of a meeting minutes management program for managing meeting minutes.

A method for generating meeting minutes using a multi-channel receiver according to an embodiment of the present invention includes at least one microphone installed individually for each meeting participant to receive the participant's voice, and a voice for the voice input from the microphone connected to the microphone. A meeting minutes generation method using a multi-channel receiver using a meeting minutes generation system using a multi-channel receiver including a multi-channel receiver connected to a user terminal to receive signals as individual channels for each microphone and transmit voice signals, wherein the user logging in by executing a meeting minutes management program installed in the terminal; individually executing, by the meeting minutes management program, voice recording with a microphone set at the beginning of a meeting; generating voice files for individually recorded voices through each microphone during the conference at the end of the conference; If there are a plurality of voice files, performing sync processing to synchronize start times of the voice files; merging the sync-processed audio files and performing STT conversion to generate one meeting transcript for one audio file; A step of generating and storing meeting minutes after text extraction by STT conversion for all voices of the voice file.

In the above, the meeting minutes management program checks whether there are a plurality of utterance sections by checking the frequency band of the voice at the same time when converting the STT for the input voice file; Separating and extracting the plurality of uttered sections into separate voices and performing STT conversion on the corresponding sections, respectively.

In the above, when all the uttered sections are checked and there is no plurality of utterance sections, the average value (RMS) of the voice frequency band is obtained to remove unnecessary noise during STT conversion from the voice file, and only the sound of the largest frequency band is taken for STT conversion It further includes the step of performing.

In the above, the meeting minutes management program further includes extracting a text result by performing STT conversion on voice collected during a set period during the meeting at the request of the user, and displaying the text result on a real-time screen.

In the above, if the meeting minutes management program that checks whether the meeting is over is during the meeting before the meeting ends, real-time displaying a text result after STT conversion for a set audio output section; Calculating an average value (RMS) of voice signals during a set time period during which voice signals are output from a plurality of microphones; performing text conversion on speech for a section that satisfies 70% or more of the maximum RMS value; The converted text is displayed on a real-time screen.

In the above, if the conference is over, a step of collectively outputting text results for voice signals recorded until the end and displaying them on a screen is further included.

The meeting minutes generation system using the multi-channel receiver of the present invention receives voice signals for each speaker through independent channels and generates voice files for the separated voice signals, so the voice recognition rate is improved and the process of separating mixed voice signals is simplified. It is unnecessary, and when extracting text from voice, there is an advantage in that text conversion for voice is performed well.

In addition, when multiple speakers speak at the same time, the existing voice-to-text conversion system that separates and extracts after simultaneous recording extracts the speaker's utterance by distinguishing only the loud sound, so it is difficult to extract text for the voice in the case of a speaker with low sound. On the other hand, in the present invention, even if a plurality of speakers speak at the same time, separately recorded voice files are generated, and when merging voice files, each voice file is classified and merged, and when converting STT, conversations for each speaker are omitted by distinguishing and merging the text. It has the advantage of being able to generate meeting minutes by extracting them without it.

In addition, the present invention provides management functions including editing and deletion of generated meeting minutes, and provides user convenience by providing a web-based meeting minutes generation and management service.

1 is a block diagram showing the overall configuration of a meeting minutes generating system using a multi-channel receiver according to an embodiment of the present invention.
FIG. 2 is a block diagram showing detailed functions of the meeting minutes management program 210 of FIG. 1 .
3 is a block diagram showing a detailed configuration of the management server 500 of FIG. 1 .
4 is a flowchart showing the entire process of a method for generating meeting minutes using a multi-channel receiver according to an embodiment of the present invention.
5A and 5B are flowcharts illustrating an audio text extraction algorithm by speaker separation in a method for generating meeting minutes using a multi-channel receiver according to an embodiment of the present invention.
6 is a flowchart illustrating a process for real-time display after audio text extraction in a method for generating meeting minutes using a multi-channel receiver according to an embodiment of the present invention.
7 is a diagram exemplarily showing the generated data format and meeting minutes format format according to an embodiment of the present invention.
8 is a diagram showing a screen displaying a list of meeting minutes notes generated by the meeting minutes management program 210 according to an embodiment of the present invention.
FIG. 9 is a view showing a screen displaying conference conversation contents by classifying attendees in chronological order when the meeting minutes are retrieved in the meeting minutes management program 210 according to an embodiment of the present invention.
10 is a diagram showing an example of a screen for displaying and changing a list of attendees in the meeting minutes management program 210 according to an embodiment of the present invention.
11 is a diagram showing an example of an attendee's microphone setting screen in the meeting minutes management program 210 according to an embodiment of the present invention.
12 is a diagram showing an example of displaying an STT conversion result on a real-time screen in the meeting minutes management program 210 according to an embodiment of the present invention.

Hereinafter, specific embodiments of the present invention will be described in detail with reference to the drawings. However, the spirit of the present invention is not limited to the presented embodiments, and those skilled in the art who understand the spirit of the present invention may add, change, delete, etc. other elements within the scope of the same spirit, through other degenerative inventions or the present invention. Other embodiments included within the scope of the inventive idea can be easily proposed, but it will also be said to be included within the scope of the inventive concept. In addition, components having the same function within the scope of the same idea appearing in the drawings of each embodiment are described using the same reference numerals.

1 is a block diagram showing the overall configuration of a meeting minutes generation system using a multi-channel receiver according to an embodiment of the present invention, and FIG. 2 is a block diagram showing detailed functions of the meeting minutes management program 210 of FIG. 1, Figure 3 is a block diagram showing the detailed configuration of the management server of Figure 1.

A meeting minutes generation system using a multi-channel receiver according to an embodiment of the present invention includes a plurality of microphones 100 and a multi-channel receiver (Blutooth) capable of collecting conversation contents of speakers independently input from the plurality of microphones 100. Multimedia Receiver: BMR, 300) and a management server that receives voice data collected from conversations of speakers through wired/wireless communication network 400, generates and provides meeting minutes, and stores and manages meeting minutes in database 550 (500) may be included.

A plurality of microphones 100 are individually installed for each conference participant to receive the participant's voice. That is, the microphone 100 can individually receive the speaker's voice and transmit the voice signal to the connected multi-channel receiver 300.

In addition, the plurality of microphones 100 may transmit audio signals through wired/wireless communication with the multi-channel receiver 300, and the communication method may be wired communication through a USB connection, Bluetooth, Wi-Fi, and the like.

The multi-channel receiver 300 is connected to the microphone 100, receives a voice signal for a voice input from the microphone 100 as an individual channel for each microphone 100, and uses the user terminal 200 to transmit the voice signal. connected with

That is, the multi-channel receiver 300 is composed of multiple channels so that separate voice files can be generated by classifying voice signals received independently from the plurality of microphones 100 for each microphone 100 .

In addition, the multi-channel receiver 300 may be connected to the management server 500 through the wired/wireless communication network 400 to transmit separate voice files to the management server 500 in order to generate meeting minutes.

Furthermore, similar to the plurality of microphones 100, the multi-channel receiver 300 may be connected to the user terminal 200 through a wired/wireless communication method such as USB connection, Bluetooth, or Wi-Fi.

The user terminal 200 may be provided to execute the meeting minutes management program 210 provided by the management server 500 and perform various functions by user input through the executed meeting minutes management program 210 .

In particular, the user terminal 200 edits or deletes some of the voice files for each speaker recorded during the meeting provided to the management server 500, and provides the modified voice files to the management server 500 to generate meeting minutes, and provides them to the user (eg, advanced level). The meeting manager) can request to generate meeting minutes with only necessary contents excluding unnecessary contents, thereby increasing user convenience by generating efficient meeting minutes.

The user terminal 200 can perform functions of creating, editing, and managing meeting minutes through the installed meeting minutes management program 210 without the help of the management server 500 through the wired/wireless communication network 400, and the management server 500 creates Only cloud storage of recorded meeting minutes, management of meeting minutes, and update functions can be provided.

In addition, referring to FIG. 2 to perform specific functions, the meeting minutes management program 210 includes a collection unit 211, an STT conversion unit 212, a generation unit 213, a management unit 214, and a server interlocking unit 215. more includes

The voice-to-text conversion (STT) is performed by receiving the merged voice file for generating meeting minutes from the multi-channel receiver 300 through the wired/wireless communication network 400, and meeting minutes can be generated using the text result.

At this time, the generated meeting minutes are created in the form of a document file that can be read by the user, and can be generated so that the conversation contents are clearly recorded for each speaker in chronological order so that the minutes can be easily checked. In conjunction with (210), it may provide meeting minutes creation, management, and editing functions.

The collection unit 211 collects voice signals corresponding to the voice signals from the multi-channel receiver 300, and transfers the voice to the STT conversion unit 212 to generate meeting minutes by outputting text for the corresponding voice.

The STT conversion unit 212 extracts text for voice data from the voice file received from the collection unit 211 .

Referring also to FIG. 7 , the format of voice data may be, for example, WAV extension, and the format of extracted text may be UTF-8.

To this end, the STT conversion unit 212 includes one or more voice recognition models, and may include a multilingual voice recognition model for supporting multiple languages, a voice recognition model for recognizing local dialects, and the like. In the case of a multilingual voice recognition model, an input voice may be automatically converted into a desired language selected by the user or a set language.

In addition, the voice recognition model can be trained by an artificial intelligence learning algorithm to improve voice recognition performance and recognize various voices.

In addition, a voice evaluation model for learning the voice of a non-native speaker and recognizing the voice may be added.

In addition, the artificial intelligence learning algorithm for improving voice recognition reflects the voice recognition result of the voice recognition model as an input variable to the learning model, so that the error is finely corrected as training is repeated, thereby improving the voice recognition rate.

Specifically, the permissible threshold range for the speech recognition result value of the speech recognition model is determined, and if the value is out of the permissible threshold range, the speech recognition rate is determined to be low, and the offset difference value is reflected to calculate the corrected speech recognition result during the next speech recognition. As the offset difference value is repeated, the correction range is reduced as the learning model is repeated, and ultimately, the voice recognition is performed within the normal voice recognition range, so that the voice recognition rate can be improved. As an artificial intelligence learning algorithm for this purpose, at least one of a support vector machine (SVM), a convolutional neural network (CNN), and a recurrent neural network (RNN), which are advantageous for pattern learning, may be used.

The generation unit 213 collects the text extracted by conversion from the STT conversion unit 212 and generates a meeting minutes file in the form of a single document.

The form of the meeting minutes file may be created in a general-purpose document format so that the contents of the meeting minutes can be deleted, edited, and corrected after viewing, and may be, for example, word format (docx) or Korean format (hwp).

The management unit 214 reads the generated meeting minutes file and provides a function of modifying and deleting the meeting minutes.

In addition, the management unit 214 may support reading and editing and deleting meeting minutes stored in the user terminal 200 or stored in the database 550 of the management server 500 according to a user's request.

Furthermore, the meeting minutes management program 210 provides various functions necessary for managing meeting minutes. Specifically, when a plurality of meeting minutes are generated, as shown in FIG. It can include editing, sorting function for time/place/user (responsible person), deletion function, voice file upload, export (save files separately), etc.

In addition, the meeting minutes management program 210 provides a user log-in function in terms of security and member management, and may include a member sign-up and log-in procedure for this purpose.

Specifically, for login authentication, membership registration is performed by applying a license key at the time of membership registration, and login may be performed based on member information registered as a member.

In addition, the license can provide a web page that can be purchased from the management server 500, and in the case of an offline environment, the user terminal to be used after storing the license key purchased offline or online in a portable storage device such as a USB memory ) and can be used after completing membership registration. In this way, after logging in as a member, the meeting minutes management program 210 provided from the management server 500 may be downloaded and installed.

In addition, referring to FIGS. 9 and 10 , the meeting minutes management program 210 may provide functions for editing, correcting, and deleting the contents of the meeting minutes, attendees (speakers), and the like, and provides an interface for a voice progress bar. Thus, it is provided to perform editing and correction while directly listening to the voice, and in particular, a function of deleting an unnecessary section, for example, a part where speech does not occur, may be included.

In addition, referring to FIG. 11, the meeting minutes management program 210 provides a setting and editing function for a plurality of microphones 100, and selects or uses a microphone 100 to be used after checking a list of the plurality of connected microphones 100. A function such as selecting a microphone 100 that has not been selected may be provided.

In addition, the meeting minutes management program 210 may use only a single microphone 100 depending on the setting of the microphone 100. This may correspond to the case where teachers, professors, lecturers, etc. are conducting classes at the pulpit. Even in this case, after generating the above-described audio file and converting the STT to the audio file, a text recording result for the audio content can be generated. Of course there is.

In addition, the meeting minutes management program 210, as an additional function, downloads the meeting minutes management program 210 initially from the management server 500 or has a server interlocking unit 215 communicating to store and manage generated meeting minutes based on the web. More may be included.

Referring to FIG. 3, the management server 500 performs cloud storage of the generated meeting minutes, management of meeting minutes, and update functions, in detail, the communication unit 510, the meeting minutes management unit 520, the statistical learning unit 530, and the app linkage. A unit 540 and a database 550 are further included.

The communication unit 510 includes at least one wired/wireless communication protocol capable of communicating with the user terminal 200 in which the meeting minutes management program 210 is installed to receive meeting minutes through wired/wireless communication.

When a plurality of user terminals 200 exist for the meeting minutes collected through the communication unit 510, the meeting minutes management unit 520 sorts them according to the sorting conditions (time, place, user, etc.) specified for each user terminal 200. It can be listed or stored in the database 550 for management.

In addition, the meeting minutes management unit 520 operates as a membership system, stores and manages member information about the user in the database 550 for security, and performs login authentication based on the member information.

In addition, the meeting minutes management unit 520 provides a license key for operating as a membership system, and allows only programs authenticated by authentication and login of the meeting minutes management program 210 installed in the user terminal 200 to be executed.

Furthermore, the meeting minutes management unit 520 replaces the offline installation type meeting minutes management program 210 installed and executed in the user terminal 200, and communicates based on a web program to generate or edit meeting minutes, and the installation type meeting minutes management program 210 and support to perform the same function.

Through this, the management server 500 may provide a web-based meeting minutes creation and editing function in real time through wired/wireless communication even if the meeting minutes management program 210 is not separately installed in the user terminal 200 .

The statistical learning unit 530 classifies meeting minutes by meeting, place, and time, stores meeting minutes in the database 550 for management, and generates statistical information on various meeting minutes to provide to users. .

The app interlocking unit 540 is installed and executed in the user terminal 200 and provides an interface compatible with the server interlocking unit 215 of the meeting minutes management program 210 for managing meeting minutes, for example, the meeting minutes management program 210. It can play a role such as receiving the edited meeting minutes and updating and storing them in the database 550.

4 is a flowchart showing the entire process of a method for generating meeting minutes using a multi-channel receiver according to an embodiment of the present invention.

In the method for generating meeting minutes using a multi-channel receiver according to an embodiment of the present invention, the meeting minutes management program 210 may be installed and executed, and a login screen for user authentication may be provided (S41).

The meeting minutes management program 210 executes a start button according to a user input when starting a meeting through setting the microphone 100 after logging in, and recording can be performed individually from each microphone 100 (S42).

Upon a user's request, the meeting minutes management program 210 extracts a text result by performing STT conversion on voice collected during a set period during the conference, and displays the text result on a real-time screen (S43).

Through this, meeting participants can check the contents of the meeting in real time through the screen, so even if they miss the contents of the conversation, they can check the contents of the conversation displayed on the screen. In particular, even in multinational conferences, since the screen is displayed by real-time translation, user convenience can be improved.

In addition, the meeting minutes management program 210 executes the meeting end button after the meeting ends to generate audio files for each speaker in an audio format.

The meeting minutes management program 210 executes a sync process for synchronizing the start time due to the delay time of the created audio file and removing voice noise, and then generates each audio file into one merged file for converting STT and generating meeting minutes ( S44).

Thereafter, the meeting minutes management program 210 extracts text for each speaker by performing STT conversion when a user requests to generate meeting minutes, and generates meeting minutes in the form of a single document format by collecting the extracted text. That is, after text extraction by STT conversion for all voices of the voice file, meeting minutes are generated (S45).

In addition, the meeting minutes management program 210 provides a service so that the created meeting minutes can be read and confirmed, and the user can directly manage the meeting minutes through a function of editing the meeting minutes and deleting some of them (S46).

At this time, it is preferable to prevent unauthorized theft or malicious editing and deletion of meeting minutes by granting a separate authority to each member to provide a limited service with limited authority for the function of editing and deleting meeting minutes.

To this end, meeting minutes may be stored and managed in the database 550 based on blockchain.

The created or edited meeting minutes may be stored in the user terminal 200 or, upon request, processed by the management server 500 for security based on the blockchain on the web, stored in the database 550, and periodically edited meeting minutes may be updated ( S47).

For blockchain-based management, a blockchain network is established in connection with multiple blockchain servers, and public and private keys are generated through the established internal blockchain network, converted into hash values, distributed and stored, and distributed and stored open. User authentication can be performed based on the key and the user's personal information.

Furthermore, each user certificate including a hash value for the user information may be generated by receiving individual user information along with a public key from the plurality of user terminals 200, and the storage method for each user certificate is a merkle tree. structure can be achieved.

For example, each user certificate (transaction) is stored with a hash value in the lowest child node, and the hash value is stored in the intermediate node on the path leading to the lowest child node in the merkle root (parent node), which is the top level of the merkle tree. It is hashed and stored so that it can be shared.

When determining the authenticity of the stored user certificate through this, the user certificate copied to the individual user terminal 200 and the user certificate in the database 550 are compared, and only the hash value hashed along the path of the Merkle tree is compared. It is done by

At this time, since the comparison operation is performed on the path of the Merkle tree, it is not necessary to perform the comparison operation on the blocks of all nodes, so it is possible to determine the authenticity with a relatively easy amount of calculation, and forgery of the transaction can be found easily and quickly. , Transactions can be easily verified even in the user terminal 200 in the form of a portable terminal with a small capacity.

5A and 5B are flowcharts illustrating an audio text extraction algorithm by speaker separation in a method for generating meeting minutes using a multi-channel receiver according to an embodiment of the present invention.

An audio text extraction algorithm by speaker separation in a method for generating meeting minutes using a multi-channel receiver according to an embodiment of the present invention performs sync processing of a voice file with respect to voices collected from a plurality of microphones 100, and simultaneous speech. It is an algorithm for separating and extracting the voice of a speaker.

Specifically, first, voice files recorded by n microphones 100 are generated according to the number of participants (n) attending the conference (S51).

The meeting minutes management program 210 checks whether the meeting has ended, and if the recording has ended, sync processing is performed on the corresponding voice files (S52 and S53).

The sync process is for synchronizing the start time and removing noise such as howling of the microphone 100 since the corresponding voice file is recorded individually and includes a slight delay in the start time.

Thereafter, the meeting record management program 210 merges n voice files into one voice file to generate one meeting record (S54).

Since the voice file includes time information on the voiced voice, the uttered voice can be checked for each time interval and converted into STT to extract the uttered content into text (S55 and S56).

That is, the meeting minutes management program 210 checks whether there are a plurality of spoken sections by checking the frequency band of the voice at the same time when converting the STT for the input voice file, and separate voices for the plurality of spoken sections. After separating and extracting, STT conversion is performed on the corresponding section (S57, S59, S60).

If there are no multiple utterance sections after checking all the utterance sections, the average value (RMS) of the voice frequency band is obtained to remove unnecessary noise (e.g., the voice of the speaker next to you) when converting STT from the voice file, and STT conversion is performed by taking only the sound (an index for identification to be extracted to the sound source of the corresponding band) is taken (S58).

When the STT conversion for all audio file sections is completed, one meeting minutes file is created in which the text is extracted and sorted in chronological order, and the content result is displayed on the user screen (S61 and S62).

6 is a flowchart illustrating a process for real-time display after audio text extraction in a method for generating meeting minutes using a multi-channel receiver according to an embodiment of the present invention.

First, when recording starts, the meeting minutes management program 210 checks whether the microphone 100 is ready for recording according to the setting of the microphone 100 (S70, S71).

After the check is completed, recording is performed with the set microphone 100 until the end of the meeting.

The meeting minutes management program 210 that has checked whether the meeting is over can provide a real-time display function of the text result after STT conversion for the set audio output section if it is during the meeting before the end of the meeting (S72).

The audio output section set here is a period in which the output time of the audio signal (sound signal) is a certain period of time, and the real-time display function converts all the audio signals of the corresponding section into text and extracts a certain amount of text for an hour period. provided in a way.

Specifically, an average value (RMS) of audio signals is calculated during a set time period during which audio signals are output from the N microphones 100 (S73).

Thereafter, text conversion is performed on speech for a section satisfying 70% or more of the maximum RMS value (S74).

The reason for converting voice text only for a section that satisfies a specific RMS value or higher is to remove unnecessary noises, such as voices from neighbors or ambient noises, which may be included in the voice collected from the corresponding microphone 100. .

The converted text is displayed on a real-time screen (S75).

Previously, if the meeting is over in step S72, text results for the voice signals recorded until the end can be collectively output and displayed on the screen and provided (S76), and an audio file for the conference is created and provided upon user request. It can be.

In this specification, a 'terminal' may be a wireless communication device that guarantees portability and mobility, and may be, for example, any type of handheld-based wireless communication device such as a smart phone, a tablet PC, or a laptop computer. In addition, the 'terminal' may also be a wired communication device such as a PC capable of accessing other terminals or servers through a communication network.

In addition, a communication network refers to a connection structure capable of exchanging information between nodes such as terminals and servers, such as a local area network (LAN), a wide area network (WAN), and the Internet (WWW : World Wide Web), wired and wireless data communications network, telephone network, and wired and wireless television communications network.

Examples of wireless data communication networks include 3G, 4G, 5G, 3rd Generation Partnership Project (3GPP), Long Term Evolution (LTE), World Interoperability for Microwave Access (WIMAX), Wi-Fi, Bluetooth communication, infrared communication, ultrasonic communication, visible light communication (VLC: Visible Light Communication), LiFi, and the like, but are not limited thereto.

100; mike
200; User terminal
210; Meeting minutes management program
211; collection department
212; STT conversion part
213; generator
214; Management
215; Server interlocking part
300; multi-channel receiver
400; wired/wireless communication network
500; management server
510; Ministry of Communications
520; Minutes management department
530; Department of Statistics
540; App interlocking department
550; database

Claims (10)

at least one microphone installed individually for each conference participant to receive the participant's voice;
a multi-channel receiver connected to the microphone and connected to the user terminal to receive a voice signal for voice input from the microphone as an individual channel for each microphone and to transmit the voice signal;
Including,
The user terminal is connected to the multi-channel receiver and has a meeting minutes management program installed and executed for generating individually recorded voice files for each received voice signal.
The meeting minutes management program,
A meeting minutes generation system using a multi-channel receiver, characterized in that it is a program for generating a meeting minutes file by performing STT conversion on the voice file, and modifying and managing the contents of the meeting minutes. According to claim 1,
The meeting minutes management program
a collecting unit that collects voice signals corresponding to the voice signals from the multi-channel receiver and transmits the collected voice to the STT conversion unit;
an STT conversion unit that includes a speech recognition model and performs text conversion on speech;
a generation unit for generating a meeting minutes file by collecting the text extracted by the conversion from the STT conversion unit;
a management unit that reads the meeting minutes file generated by the generating unit and provides a function of modifying or deleting the meeting minutes upon a user's request;
Meeting minutes generation system using a multi-channel receiver comprising a. According to claim 2,
Server interlocking unit that initially downloads a meeting minutes management program from a management server that communicates with user terminals through a wired/wireless communication network, or stores and manages generated meeting minutes in a web-based manner.
Meeting minutes generation system using a multi-channel receiver further comprising a. According to claim 3,
The management server
a communication unit capable of communicating with a user terminal in which the meeting minutes management program is installed to receive the meeting minutes through wired/wireless communication;
In order to store and manage the meeting minutes collected through the communication unit in the database, member information is provided and stored to operate as a membership system, log-in authentication for the meeting minutes management program and database storage function is performed based on the member information, and the meeting minutes management A meeting minutes management unit that supports a function of generating or editing meeting minutes by communicating based on a web program on behalf of the program;
a statistical learning unit that classifies the meeting minutes by meeting, place, and time, stores and manages the minutes in a database, and generates and provides statistical information on various meeting minutes to users;
App interlocking unit that is installed and executed in the user terminal and provides an interface compatible with the server interlocking unit of the meeting minutes management program for managing meeting minutes
Meeting minutes generation system using a multi-channel receiver further comprising a. At least one microphone that is installed individually for each conference participant to receive the participant's voice, and is connected to the microphone to receive a voice signal for the voice input from the microphone as an individual channel for each microphone, and to transmit the voice signal to the user In the meeting minutes generation method using a multi-channel receiver using a meeting minutes generation system using a multi-channel receiver including a multi-channel receiver connected to a terminal,
logging in by executing a meeting minutes management program installed in the user terminal;
individually executing, by the meeting minutes management program, voice recording with a microphone set at the beginning of a meeting;
generating voice files for individually recorded voices through each microphone during the conference at the end of the conference;
If there are a plurality of voice files, performing sync processing to synchronize start times of the voice files;
merging the sync-processed audio files and performing STT conversion to generate one meeting transcript for one audio file;
Generating and storing meeting minutes after text extraction by STT conversion for all voices of the voice file;
Meeting minutes generation method using a multi-channel receiver comprising a. According to claim 5,
The meeting minutes management program checks whether there are a plurality of utterance sections by checking the frequency band of the voice at the same time when converting the STT for the input voice file;
Separating and extracting the plurality of uttered sections into separate voices and performing STT conversion on each of the corresponding sections;
Meeting minutes generation method using a multi-channel receiver further comprising. According to claim 6,
If there is no plurality of utterance sections by checking all the utterance sections, obtaining an average value (RMS) of the voice frequency band to remove unnecessary noise during STT conversion from the voice file and taking only the sound of the largest frequency band to perform STT conversion ;
Meeting minutes generation method using a multi-channel receiver further comprising. According to claim 5,
The meeting minutes management program extracts a text result by performing STT conversion on voice collected during a set period of time at the user's request, and displaying the text result on a real-time screen;
Meeting minutes generation method using a multi-channel receiver further comprising. According to claim 8,
If the meeting minutes management program that has checked whether the meeting is over is in the middle of the meeting before the end of the meeting, real-time displaying a text result after converting the STT for the set audio output section;
Calculating an average value (RMS) of voice signals during a set time period during which voice signals are output from a plurality of microphones;
performing text conversion on speech for a section that satisfies 70% or more of the maximum RMS value;
displaying the converted text on a real-time screen;
Meeting minutes generation method using a multi-channel receiver further comprising. According to claim 9,
If the meeting is over, outputting the text results of the voice signals recorded until the end of the meeting in batches and displaying them on the screen to provide them
Meeting minutes generation method using a multi-channel receiver further comprising.

Images