CN105810208A - Meeting recording device and method thereof for automatically generating meeting record - Google Patents

Abstract

The invention provides a meeting recording device and a method thereof for automatically generating a meeting record. The method comprises the steps of recognizing a silent fragment in speech data; judging whether the time of the silent fragment is greater than a preset value or not; taking the silent fragment of which the time is greater than the preset value as the boundary, segmenting the speech data or characters obtained by converting the speech data; and generating an original meeting record according to the segmentation condition of the speech data or the characters and a meeting recording template. The meeting recording device and the method thereof for automatically generating a meeting record can automatically generate the meeting record according to the preset meeting recording template, and are thus more time-saving, convenient and humanized than the prior art.

Description

Conference recording device and method for automatically generating conference records

technical field

The invention relates to a conference record device and a method for automatically generating conference records.

Background technique

The existing method for reporting and recording in a meeting usually uses equipment such as a video camera, a microphone, and a recording pen to record and video the speeches of each person during the meeting. Those who make meeting minutes after the meeting can view and play back audio and video recordings to organize meeting minutes. However, manually marking and extracting voice data is time-consuming and extremely inconvenient for users.

Contents of the invention

In view of this, it is necessary to provide a conference record device and a method for automatically generating conference records, which can automatically generate conference records to solve the above problems.

The invention provides a conference recording device, which includes a memory and a processor. The meeting recording device also includes the following modules controlled by the processor and stored in the memory: an identification module, used to identify the silent segment in the voice data; a judging module, used to judge the silent segment experienced Whether the time is greater than a preset value; the segmentation module is used to segment the voice data or the converted text of the voice data with the silent segment whose elapsed time is greater than the preset value; and a generation module is used to generate an original meeting record according to the situation where the voice data or the text is segmented and the meeting record template stored in the memory.

The present invention also provides a method for automatically generating meeting minutes, which runs in at least one device including a memory and a processor. The method includes the following steps of being executed by the processor controlling the modules stored in the memory: identifying step: identifying the silent segment in the voice data; judging step: judging whether the elapsed time of the silent segment is greater than a preset Value; Segmentation step: with the silent segment whose elapsed time is greater than the preset value as the boundary, the speech data or the converted text of the speech data is segmented; and generation step: according to the speech data or the The situation that the text is segmented and the meeting record template stored in the memory generate an original meeting record.

The meeting record device and the method for automatically generating meeting minutes described in the present invention can automatically generate meeting minutes according to a preset meeting record template, thus, compared with the existing methods, it is more time-saving, convenient and humanized.

Description of drawings

FIG. 1 is a schematic diagram of an application environment of a meeting recording device according to an embodiment of the present invention.

FIG. 2 is a functional block diagram of an embodiment of the conference recording device shown in FIG. 1 .

Fig. 3 is a schematic diagram of the generated original meeting minutes and the edited meeting minutes in one embodiment of the present invention.

4-7 are flow charts of the steps of the method for automatically generating meeting minutes in different embodiments of the present invention, respectively.

Description of main component symbols

conference recording device 100 cloud device 200 user 1 original meeting minutes 310 Edited Minutes 320 How to Automatically Generate Meeting Minutes 400, 500, 600, 700 memory 10 recording module 11 conversion module 12 Identification module 13 judgment module 14 Proofreading and editing module 15 build module 16 sending module 17 Segmentation module 18 control module 19 voice input unit 20 touch screen 30 communication unit 40 positioning module 50 processor 60 step S401-S407, S501-S508, S601-S607, S701-S707

The following specific embodiments will further illustrate the present invention in conjunction with the above-mentioned drawings.

detailed description

Please refer to FIG. 1 , which is a schematic diagram of an application environment of a conference recording device 100 according to an embodiment of the present invention. In this embodiment, the conference recording device 100 can be connected with a cloud device 200 . Wherein, the meeting recording device 100 is located near each user 1 and can receive the voice of each user 1 in the meeting or report, that is, the speech of the user 1 . The meeting recording device 100 and/or the cloud device 200 has the function of automatically generating meeting minutes according to the voice received by the meeting recording device 100 . User 1 is a participant in a meeting or report. For convenience of description, meetings or reports are collectively referred to as meetings below.

In one embodiment, the meeting recording device 100 has a function of automatically generating meeting minutes, that is, it can generate meeting minutes by itself. Moreover, the conference recording device 100 does not rely on the cloud device 200 , but automatically generates conference records according to the voice it receives. When multiple users 1 hold meetings or reports, the meeting recorder 100 can automatically record the voices of each user 1, and automatically recognize the voices of each user, and convert the recognized voices into text, and record them according to the preset meeting records. The template automatically generates meeting minutes and automatically sends them to relevant personnel according to the preset method. Relevant personnel include each user 1 and/or other relevant personnel of the meeting, such as the person in charge of the to-do items, relevant supervisors, and the like. So as to realize the function of automatically recording, generating and sending meeting minutes.

For the convenience of description, the words in the following brackets in this paragraph are simplified functional descriptions of the preceding words. For details, please refer to the following instructions. The meeting recording device 100 can automatically identify each corresponding user of the received voice (identify the user in the voice), and then convert the received voice into text including the user name of the recognized user, or automatically convert the received voice into text (convert voice to text), and then recognize the user name of each user 1 from the text (identify the user in the text). Then divide the text into paragraphs according to the above-mentioned user names recognized from the voice and/or text (divide paragraphs according to the text), and then automatically generate meeting minutes according to the preset meeting record template (generate meeting minutes). The meeting recorder 100 can also automatically recognize the silent segment in the received voice (identify the silent segment according to the voice), divide the speech into multiple voice segments according to the recognized silent segment (divide the paragraphs according to the voice), and then divide the multiple voice segments into segments respectively. Convert a speech clip into corresponding text (speech to text), and then automatically generate meeting minutes according to the preset meeting record template (generating meeting minutes). The meeting recording device 100 can also automatically recognize words and sentences that appear repeatedly in voice and/or text information, and store them in the database of commonly used words. words and sentences.

In another embodiment, the conference recording device 100 can perform data communication with the cloud device 200, so that the conference recording device 100 and the cloud device 200 together or the cloud device 200 alone automatically generates conference records according to the voice received by the conference recording device 100. Therefore, in the present invention, the conference recording device 100 can also record the conference, and convert the recorded voice into a voice signal, and convert the converted voice signal and/or other data (such as text converted from the voice signal, etc.) ) to the cloud device 200, and the conference recording device 100 and/or and the cloud device 200 respectively perform all or part of the following functions performed by the conference recording device 100 in the previous embodiment: voice conversion into text, Recognize users in voice and/or text, identify silent segments based on voice and/or text, divide paragraphs based on voice and/or text, generate meeting minutes, recognize common words and sentences in voice and/or text, store common words and sentences in frequently used language database, and automatically proofread/edit text or meeting minutes based on commonly used words and sentences.

Please refer to FIG. 2 , which is an embodiment of the present invention. It should be noted that, what is shown in FIG. 2 is only a functional block diagram of the meeting recording device 100 in one embodiment of the present invention. Corresponding to the various embodiments of the present invention described above, the meeting recording device 100 may also only include Figure 2 shows a part of the functional units/modules. The cloud device 200 may include other functional units/modules shown in FIG. 2 . For example, in an implementation in which the function of automatically generating conference minutes is performed solely by the cloud device 200, the conference recording device 100 may include the voice input unit 20, the communication unit 40, and the processor 60 shown in FIG. 2, and the cloud device 200 may include corresponding The communication unit, the processor and the modules 12-19 stored in the memory 10. Corresponding descriptions will be made below when necessary.

In this embodiment, the conference recording device 100 includes a memory 10 , a voice input unit 20 , a touch screen 30 , a communication unit 40 , a positioning module 50 and a processor 60 . The memory 10, voice input unit 20, touch screen 30, and communication unit 40 are respectively connected to the processor 60 through signal lines and data lines. The conference recording device 100 is a smart phone. In other embodiments, the conference recording device 100 may also be a tablet computer, a notebook computer, a desktop computer, a conference phone, and the like.

In this embodiment, the meeting recording device 100 can independently and automatically generate meeting minutes. The meeting recording device 100 automatically converts the received voice into text according to the voice received by the voice input unit 20 of the user 1 participating in the meeting, and then automatically generates a meeting record according to a preset meeting record template. Specifically, the conference recorder 100 can perform the aforementioned automatic conversion of the received voice into text, automatically identify users in the received voice or the converted text, divide the text into paragraphs according to the recognized user name, and then The meeting minutes template automatically generates meeting minutes. The meeting recorder 100 can also automatically recognize the silent segment in the received voice, divide the voice into multiple voice segments according to the recognized silent segment, and then convert the multiple voice segments into corresponding text, and then according to the preset The meeting minutes template automatically generates meeting minutes. The meeting recording device 100 can also automatically recognize words and sentences that appear repeatedly in voice and/or text information, and store them in the database of commonly used words. words and sentences. The meeting recording device 100 can also automatically send the generated meeting minutes and/or to-do items to the communication addresses of relevant personnel according to a preset method. Wherein, the preset manner includes a preset sending format, a preset sending time, and the like. The correspondence address of the relevant person includes at least one of the following: email address, phone number, social account (such as QQ number, WeChat account), etc.

A user voice feature table is stored in the memory 10, and the voice feature table records the one-to-one correspondence between multiple user names and their voice feature parameters. In this embodiment, the user name may be the real name of the user, or may be a nickname or a code name. The user voice feature table can be obtained through pre-training, that is, it is obtained by performing voice training and collection on each user within a period of time before the start of the meeting/report. The memory 10 may also store a meeting record template preset by the user or the system. The memory 10 can also be used to store recorded speech data, a speech-to-text database required for speech-to-text conversion, and a database of commonly used words. Wherein, the database of commonly used words is accumulated, screened and stored during the process of the conference recorder 100 executing its function of automatically generating conference minutes, or it may be downloaded and stored from a database of commonly used words.

In this embodiment, the voice input unit 20 is used to collect the voices of each user during the conference, and convert the collected voices into voice signals. The voice input unit 20 is a microphone. The communication unit 40 is used for performing data communication with the cloud device 200 in response to the control of the processor 60 . The positioning module 50 is used to provide real-time location information of the conference recording device 100 , which may be a GPS positioning module.

In one embodiment, the conference recording device 100 further includes a touch screen 30 .

In this embodiment, the memory 10 also stores a plurality of functional modules configured to be executed by one or more processors (one processor 60 in this embodiment), so as to complete the present invention. For example, as shown in FIG. 1, a recording module 11, a conversion module 12, a recognition module 13, a judgment module 14, a proofreading and editing module 15, a generation module 16, a sending module 17, a segmentation module 18 and a control module 19 are stored in the memory 10. In other embodiments, the functional modules stored in the memory 10 can also be changed according to actual needs, for example, when the voice is converted into text, the common words and sentences in the voice and/or text are automatically recognized, and the common words and sentences are stored in the common language database. , and when one or more functions in functions such as automatic text proofreading according to commonly used words and phrases are set to be executed by the cloud device 200, the memory 10 of the meeting recording device 100 may not store the functional modules required to perform the functions. The module referred to in the present invention is a program segment that completes a specific function, and is more suitable for describing the execution process of software in the processor 60 than a program. The functions of each module will be described in detail in the flowcharts of Fig. 4-Fig. 7 .

It should be noted that, for the convenience of description, in the following introductions about the method for automatically generating meeting minutes, this method is used to run in a meeting recording device (such as the meeting recording device 100) including corresponding units and/or functional modules for the introduction. According to the previous introduction, it can be seen that in the following methods for automatically generating meeting minutes, some steps can also be set to be executed by a cloud device (such as cloud device 200) connected to the meeting recording device. Therefore, correspondingly, when necessary, The steps of the conference recording device transmitting voice signals/data, text data and/or other data to the cloud device, and the cloud device receiving the signals/data can be added to the steps of the following methods for automatically generating conference minutes. Since these are some technical means that can be implemented by those skilled in the art based on the content disclosed in this specification, for the sake of saving space, they will not be described in detail in this specification one by one.

As shown in FIG. 4 , it is a flowchart of a method 400 for automatically generating meeting minutes according to an embodiment of the present invention. The method 400 for automatically generating conference minutes is to run on the conference recording device and/or the cloud device after the meeting recording function of the conference recording device (such as the conference recording device 100) and/or the cloud device (such as the cloud device 200) is enabled. Yes, it may start from step S401, step S402 or step S403.

Step S401, receiving step: the voice input unit 20 receives voice and converts the received voice into a corresponding voice signal. In this embodiment, the meeting recorder 100 is set near the user 1 in the meeting, and the voice input unit 20 is a microphone set in the meeting recorder 100 .

In another embodiment, the following steps may be executed simultaneously or before this step S401: the control module 19 controls to start the positioning module 50 to obtain the location information of a meeting recorder 100 and the current meeting time information, and the acquired The location information and time information are stored in the memory 10 . In other implementation manners, the conference recording device 100 may also receive and store information related to the current conference input via the touch screen 30 , for example, the date, time, location of the conference, and names of persons attending the conference.

Step S402 , recording step: the recording module 11 records the voice signal into voice data, and stores the recorded voice data in the memory 10 . In one embodiment, in response to the user's selection, this step may also be omitted, and step S403 is directly executed.

Step S403 , identification step: the identification module 13 identifies one or more users corresponding to the voice signal according to the voice signal and the user voice feature table stored in the memory 10 . In this embodiment, the recognition module 13 obtains one or more speech features according to the speech signal analysis, and queries one or more users corresponding to the same/most similar speech features from the speech feature table, thereby obtaining the speech The corresponding one or more users in the data. When a meeting or a report is in progress, when multiple users speak/speak, the identification module 13 can identify which user's voice is included in the voice data according to the voice signal and the voice feature table.

In another embodiment, the identification module 13 also adds different tags to the voice segments of different users, and adds the same tag to the voice segments of the same user.

Step S404, conversion step: the conversion module 12 converts the voice signal into text containing the usernames of the one or more users. In this embodiment, the conversion module 12 converts the speech signal into text according to the speech signal and the speech-text database stored in the memory 10, and recognizes the speech of each user of one or more users recognized by the recognition module 13 The user name of the corresponding user is automatically added to a preset position of the converted text corresponding to the signal. In this embodiment, the preset position is the front end of the converted text corresponding to each user's voice signal.

In another embodiment, when the recognition module 13 adds some tags to speech segments of different users, the text converted by the conversion module 12 also includes these tags.

Step S405 , generating step: the generating module 16 generates an original meeting record according to the converted text and the meeting record template stored in the memory 10 . Please refer to FIG. 3 , which shows an original meeting record 310 generated by the generation module 16 in one embodiment.

In one embodiment, the generating module 16 also automatically adds the location information and time information acquired by the positioning module 50 to the generated original meeting minutes. For example, time information is added to the field of meeting date/time in the meeting minutes template, location information is added to the field of meeting location in the meeting minutes template, and so on.

The generating module 16 can also automatically add the meeting participants/attendees input by the user through the touch screen 30 to the attendee/attendee column in the meeting record template.

In another embodiment, the generation module 16 may also use the user name contained in the text recognized by the recognition module 13 or the user name of the user who uttered the voice corresponding to the voice signal recognized by the recognition module 13 according to the voice signal , to automatically add those usernames to the Attendees/Attendees field in the meeting minutes template.

Step S406 , proofreading and editing step: the proofreading and editing module 15 proofreads and/or edits the original meeting record according to preset proofreading and editing rules to obtain a meeting record.

In this embodiment, the preset proofreading and editing rule is to divide the text into paragraphs from each user name in the text. The recognition module 13 also recognizes/recognizes the user name of the user from the converted text, and the proofreading and editing module 15 divides the original meeting record into paragraphs according to the user name contained in the text recognized by the recognition module 13 . For example, the proof-editing module 15 divides paragraphs by the first or last letter of the username. When the text contains the user name Wang Daming, the proofreading and editing module 15 then uses the three texts of Wang Daming as the beginning of the paragraph. It should be noted that, in this embodiment, preferably, the user name mentioned here is the user name of the user obtained by the recognition module 13 through speech recognition. In another embodiment, the user names can also be automatically recognized by the recognition module 13 from the text according to the user names previously stored in the memory 10 . Please refer to FIG. 3 , which shows an edited meeting record 320 obtained after the proofreading and editing module 15 proofreads and/or edits the original meeting record 310 in one embodiment.

In another embodiment, the preset proofreading and editing rule is to segment the text paragraphs from the text corresponding to the beginning of each voice segment according to the tags added by the recognition module 13 to the voice segments of different users .

In yet another embodiment, the proofreading and editing module 15 also stores the proofreading and editing of the meeting minutes in the memory 10 . Alternatively, the sending module 17 controls to send the proofread and edited meeting minutes to the cloud device 200 through the communication unit 40 , so as to control to store the meeting minutes in the cloud device 200 .

In other embodiments, the proofreading and editing module 15 also edits the meeting minutes according to the editing signal generated by the touch screen 30 . For example, the user can input editing content and/or editing operations on the original meeting record through the touch screen 30, thereby providing a function for the user to manually edit the original meeting record. In addition, the preset proofreading and editing rules also include intelligent identification of proofreading text, etc., please refer to the following description based on FIG. 5 for details.

Step S407, sending step: the sending module 17 automatically sends the proofread and/or edited meeting minutes to the communication addresses of the meeting related personnel according to the preset sending rules. In this embodiment, the preset sending rule can be sent immediately (that is, sent immediately after the meeting record is generated) to the communication address of the relevant person in the meeting, or can be sent at a preset time after the meeting record is generated. The mailing address of the person involved in the meeting. The people involved in the meeting may include one or more of the following: meeting attendees, users whose usernames appear in the meeting minutes, users involved/mentioned in the meeting minutes (for example, users with to-do items), Default supervisor, responsible person, responsible person, etc.

In another embodiment, the preset sending rule may also include sending the generated meeting minutes to the mailing address of the person related to the to-do item within a preset number of days before the preset due date of the to-do item, for example , which may include the person directly responsible for the to-do item, the relevant supervisor, and other relevant personnel related to the to-do item.

In other implementation manners, this step S407 may not be set, and the user directly manually sends the meeting record to the communication address of the meeting related personnel; or, when the cloud device 200 receives and stores the meeting record, the cloud device 200 will send the meeting record to The meeting minutes are sent to the relevant personnel of the meeting.

As shown in FIG. 5 , it is a flowchart of a method 500 for automatically generating meeting minutes according to an embodiment of the present invention. The method 500 for automatically generating meeting minutes is run on a meeting recording device (such as the meeting recording device 100 ) after the meeting recording function is turned on. It should be noted that some of the steps performed in the method 500 for automatically generating meeting minutes shown in FIG. 5 and the method 400 for automatically generating meeting minutes shown in FIG. When describing the method 400 for automatically generating meeting minutes in , some alternative implementations that can be executed simultaneously for a certain step are also applicable to the same or similar ones in the method 500 for automatically generating meeting minutes in FIG. 5 The steps will not be repeated here. The method 500 for automatically generating meeting minutes may start at step S501.

Step S501, receiving step: the voice input unit 20 receives voice and converts the received voice into a corresponding voice signal.

Step S502 , recording step: the recording module 11 records the voice signal into voice data, and stores the recorded voice data in the memory 10 . In one embodiment, in response to the user's selection, this step can also be omitted, and step S503 is directly executed.

Step S503, identifying step: the identifying module 13 identifies the silent segment in the voice data according to the voice signal. In this implementation manner, the silent segment is a segment of mute data in the voice data, that is, a segment of mute in the voice. For example, when the volume of a voice segment of voice data corresponding to a certain part of the voice signal is smaller than a preset silence threshold, the identification module 13 identifies the voice segment as a silent segment. The voice data may contain multiple silent segments.

In one embodiment, when step S502 is not included, in this step, the recognition module 13 recognizes a silent segment in the speech according to the speech signal.

Step S504, judging step: the judging module 14 judges whether the elapsed time of the silent segment is greater than a preset value, if yes, execute step S505, otherwise, the process ends. In one embodiment, the preset value is 3 seconds.

Step S505, dividing step: the dividing module 18 divides the voice data into multiple voice data segments according to the silent segment. In this embodiment, the segmentation module 18 divides the voice data from the silent segment, and when the voice data includes multiple silent segments whose elapsed time is greater than the preset value, the segmentation module 18 The plurality of silent segments divides the voice data into a plurality of voice data segments.

Step S506, identification step: the identification module 13 identifies one or more users corresponding to the plurality of voice data segments according to the voice signals corresponding to the multiple voice data segments obtained by segmentation and the user voice feature table stored in the memory 10 . In an implementation manner, the method 500 for automatically generating meeting minutes may not include this step.

Step S507, conversion step: the conversion module 12 converts the voice signals corresponding to the segmented multiple voice data segments into texts including multiple paragraphs. In this embodiment, the conversion module 12 converts the plurality of voice data segments to the corresponding voice signal according to the voice signals corresponding to the multiple voice data segments, one or more users identified by the recognition module 13, and the voice and text database stored in the memory 10. The corresponding voice signal is converted into text including a plurality of paragraphs corresponding to each voice data segment.

Step S508 , generating step: the generating module 16 generates an original meeting record according to the converted text containing multiple paragraphs and the meeting record template stored in the memory 10 . The specific manner of this step S509 may be the same as the method 400 for automatically generating meeting minutes, and will not be repeated here.

In this embodiment, step S406 (proofreading and editing step) and step S407 (sending step) in the method 400 for automatically generating meeting minutes may be performed after step S508 , and details will not be repeated here.

As shown in FIG. 6 , it is a flowchart of a method 600 for automatically generating meeting minutes according to an embodiment of the present invention. The method 600 for automatically generating meeting minutes is run on a meeting recording device (such as the meeting recording device 100 ) after the meeting recording function is turned on. It should be noted that some of the steps performed in the method 600 for automatically generating meeting minutes shown in FIG. 6 and the methods for automatically generating meeting minutes shown in FIGS. 5 and 4 are the same or similar. Therefore, When describing the method 400 for automatically generating meeting minutes in FIG. 4 and the method 500 for automatically generating meeting minutes in FIG. 5 , some alternative implementations that can be executed simultaneously for a certain step are also applicable to The same or similar steps in the method 600 for automatically generating meeting minutes in FIG. 6 will not be repeated here. The method 600 for automatically generating meeting minutes may start at step S601.

Step S601, receiving step: the voice input unit 20 receives voice and converts the received voice into a corresponding voice signal.

Step S602 , recording step: the recording module 11 records the voice signal into voice data including a recording time stamp, and stores the recorded voice data in the memory 10 . In one embodiment, in response to the user's selection, this step can also be omitted, and step S603 is directly executed.

Step S603, identification step: the identification module 13 identifies one or more users corresponding to the voice signal according to the voice signal and the user voice feature table stored in the memory 10 . In one embodiment, the identification module 13 identifies one or more users corresponding to the voice signal according to the voice data including the recording time stamp and the user voice feature table stored in the memory 10 . In another implementation manner, the method 600 for automatically generating meeting minutes may not include this step.

Step S604, conversion step: the conversion module 12 converts the voice signal into text including the recording time stamp and the user names of the one or more users. In this embodiment, the conversion module 12 converts the voice signal into text including the recording time stamp and the user names of the one or more users. The conversion module 12 converts the voice signal according to the voice signal, the voice data recorded by the recording module 11 that includes the recording time stamp, one or more users recognized by the recognition module 13, and the voice text database stored in the memory 10. It is converted into text including the recording time stamp, and the corresponding user's username is automatically added to the front of the text obtained by converting the voice signal of each user. In another embodiment, the conversion module 12 converts the speech signal into a speech data containing the time stamp according to the speech signal, the speech data recorded by the recording module 11 and the speech and text database stored in the memory 10. Text describing the timestamp of the recording.

Step S605, judging step: the judging module 14 judges whether the time interval recorded in the recording time stamp corresponding to the adjacent text reaches a preset value according to the converted text, if yes, execute step S606, otherwise, The process ends. In one embodiment, the preset value is 3 seconds. The adjacent text containing the recording time stamp may contain multiple time intervals reaching the preset value.

Step S606, segmenting step: the segmenting module 18 divides text into paragraphs by dividing the adjacent text whose time interval recorded in the corresponding recording time stamp reaches the preset value. In this embodiment, specifically, the adjacent text is divided into the previous paragraph and the next adjacent paragraph, until the time intervals recorded in all corresponding recording time stamps reach the preset value of each phase. Adjacent text is divided into different paragraphs.

Step S607, generating step: the generating module 16 generates an original meeting record according to the text after the paragraphs are divided and the meeting record template stored in the memory 10 . The specific manner of this step S607 may be the same as the method 500 for automatically generating meeting minutes, and will not be repeated here.

As shown in FIG. 7 , it is a flowchart of a method 700 for automatically generating meeting minutes according to an embodiment of the present invention. The method 700 for automatically generating meeting minutes is executed on a meeting recording device (such as the meeting recording device 100 ) after the meeting recording function is turned on. It should be noted that some of the steps performed in the method 700 for automatically generating meeting minutes shown in FIG. 7 and the methods for automatically generating meeting minutes shown in FIGS. 5 and 4 are the same or similar. Therefore, When describing the method 400 for automatically generating meeting minutes in FIG. 4 and the method 500 for automatically generating meeting minutes in FIG. 5 , some alternative implementations that can be executed simultaneously for a certain step are also applicable to The same or similar steps in the method 700 for automatically generating meeting minutes in FIG. 7 will not be repeated here. The method 700 for automatically generating meeting minutes may start at step S701.

Step S701 , database building step: the control module 19 creates a database of commonly used expressions including common expressions and correction objects, and stores the database of commonly used expressions in the memory 10 . In this embodiment, when the conference recording device 100 uses the function of automatically generating conference records for the first time, the control module 19 automatically establishes the common phrase database. The common phrase database includes at least one common phrase and its corresponding relationship with its correction object, and each common phrase corresponds to at least one correction object. The common words include one or more of the following: common words, common words, common sentences, etc., and may also be voice data or text data. The correction object of each commonly used term can be accumulated and recorded during the process of manual editing and modification of meeting minutes by the user. The correction object includes one or more of the following speech data and/or text data: characters, words, sentences, etc.

In another implementation manner, the method 700 for automatically generating meeting minutes may not include this step S701. Instead, a database of commonly used words is pre-stored in the meeting recording device, and the database of commonly used words is accumulated, screened and stored during the process of the meeting recording device 100 performing its function of automatically generating meeting records, or it can be obtained from a database of commonly used words. downloaded and stored in .

Step S702, receiving step: the voice input unit 20 receives the voice and converts the received voice into a corresponding voice signal.

Step S703, conversion step: the conversion module 12 converts the voice signal into text. In one embodiment, other steps between the receiving step and the transforming step included in any one of the methods 400, 500 and 600 for automatically generating meeting minutes may also be included. That is, the step of converting the voice signal into text in the various embodiments described above may be included.

Step S704, step of identifying and storing commonly used words: when the judging module 14 recognizes and judges that the voice data and/or the text contain words and sentences that repeat a preset number of times, use the words and sentences that repeat the preset number of times as Common words are stored in the common language database. In this embodiment, the words and sentences that appear repeatedly for the preset number of times may be voice data and/or text data such as words, phrases, and sentences. In an implementation manner, this step S704 may also be omitted. The preset number of times is 20 times.

Step S705, judging step: the judging module 14 judges whether the converted text contains a correction object, if yes, execute step S706, otherwise, the process ends.

Step S706, proofreading step: the proofreading and editing module 15 automatically corrects the correction objects contained in the text into corresponding common phrases according to the common phrase database. In an embodiment, this step S706 may also be performed after step S707.

Step S707 , generating step: the generating module 16 generates an original meeting record according to the corrected text and the meeting record template stored in the memory 10 . The specific manner of this step S707 may be the same as the method 500 for automatically generating meeting minutes, and will not be repeated here.

The meeting recording device 100 and the method for automatically generating meeting minutes provided by the present invention can automatically generate meeting minutes according to a preset meeting record template, and can perform intelligent speech and text recognition, content formatting, editing and proofreading on the meeting records. Moreover, meeting minutes can also be sent to relevant personnel according to preset rules. Therefore, compared with the existing methods, it is more time-saving, convenient and humanized.

Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention without limitation. Although the present invention has been described in detail with reference to the preferred embodiments, those of ordinary skill in the art should understand that the technical solutions of the present invention can be Modifications or equivalent replacements can be made without departing from the spirit and scope of the technical solutions of the present invention.

Claims (12)

1. A method for automatically generating meeting minutes, running in at least one device comprising a memory and a processor, characterized in that the method comprises the following steps of being executed by a module stored in the memory controlled by the processor: Recognition step: recognize the silent segment in the speech data; Judging step: judging whether the elapsed time of the silent segment is greater than a preset value; Segmentation step: Segment the speech data or the text obtained by converting the speech data, taking the silent segment whose elapsed time is greater than the preset value as the boundary; and Generating step: generating an original conference record according to the speech data or the situation of the text being segmented and the conference record template stored in the memory. 2. The method according to claim 1, further comprising an editing step: editing the original meeting minutes according to preset proofreading and editing rules to obtain a meeting record. 3. The method of claim 1, wherein: The identifying step is: identifying the silent segment in the voice data according to the voice signal corresponding to the voice data; The dividing step is: when it is judged that the elapsed time of the silent segment is greater than a preset value, dividing the voice data into a plurality of voice data segments according to the silent segment; After the segmentation step, a conversion step is also included: converting the voice signals corresponding to the segmented multiple voice data segments into text comprising multiple paragraphs; The generating step is: generating an original meeting record according to the converted text containing multiple paragraphs and the meeting record template stored in the memory. 4. The method according to claim 3, further comprising an identification step: according to the voice signals corresponding to a plurality of voice data segments obtained by segmentation and the user voice feature table stored in the memory, identify the one or more users corresponding to the plurality of voice data segments; and The conversion steps include: According to the voice signals corresponding to the multiple voice data segments and the voice-text database stored in the memory, convert the voice signals corresponding to the multiple voice data segments into a plurality of paragraphs corresponding to each voice data segment one-to-one the text of; and The user name of the corresponding user is automatically added to a preset position of the text corresponding to the voice signal of the one or more users. 5. The method according to claim 3 or 4, wherein: the silent segment is a segment of silent data in the voice data; the method also includes a recording step: according to the corresponding voice of the voice data The signal records voice data, and stores the recorded voice data in a memory. 6. The method according to claim 1 or 2, further comprising: Recording step: recording voice data according to a voice signal corresponding to the voice data; the recorded voice data includes a recording time stamp; Converting step: converting the voice signal into text containing the recording time stamp; The judging step is: according to the converted text, it is judged whether the time interval recorded in the recording time stamp corresponding to the adjacent text reaches the preset value; The segmentation step is: when the time interval recorded in the recording time stamp corresponding to the adjacent text reaches the preset value, divide the time interval recorded in the corresponding recording time stamp into the preset value. The adjacent text is bounded to demarcate paragraphs of text. 7. A meeting recording device, comprising a memory and a processor, characterized in that it also includes the following modules controlled by the processor and stored in the memory: An identification module, configured to identify silent segments in the speech data; A judging module, configured to judge whether the elapsed time of the silent segment is greater than a preset value; A segmentation module, configured to segment the speech data or the text obtained by converting the speech data with the silent segment whose elapsed time is greater than the preset value as a boundary; and A generating module, configured to generate an original meeting record according to the speech data or the situation in which the text is segmented and the meeting record template stored in the memory. 8. The meeting record device according to claim 7, further comprising a proofreading and editing module, configured to edit the original meeting record according to preset proofreading and editing rules to obtain a meeting record. 9. The conference recording device according to claim 7, characterized in that: The identification module recognizes the silent segment in the voice data according to the voice signal corresponding to the voice data; When the segmentation module judges that the elapsed time of the silent segment is greater than a preset value, the voice data is divided into a plurality of voice data segments according to the silent segment; The meeting recording device also includes a conversion module, which is used to convert the voice signals corresponding to the segmented multiple voice data segments into text comprising multiple paragraphs; The generating module generates an original meeting record according to the converted text containing multiple paragraphs and the meeting record template stored in the memory. 10. The conference recording device according to claim 9, characterized in that: The identification module is also used to identify one or more users corresponding to the multiple voice data segments according to the voice signals corresponding to the multiple voice data segments obtained through segmentation and the user voice feature table stored in the memory; The conversion module is also used to: According to the voice signals corresponding to the multiple voice data segments and the voice-text database stored in the memory, convert the voice signals corresponding to the multiple voice data segments into a plurality of paragraphs corresponding to each voice data segment one-to-one the text of; and The user name of the corresponding user is automatically added to a preset position of the text corresponding to the voice signal of the one or more users. 11. The conference recording device according to claim 9 or 10, wherein: the silent segment is a segment of silent data in the voice data; the conference recording device also includes a recording module for record the voice data corresponding to the voice signal corresponding to the voice data, and store the recorded voice data in the memory. 12. The conference recording device according to claim 7 or 8, further comprising a recording module and a conversion module, wherein: The recording module is used to record voice data according to a voice signal corresponding to the voice data; the recorded voice data includes a recording time stamp; The conversion module is used to convert the voice signal into text containing the recording time stamp; The judgment module judges whether the time interval recorded in the recording time stamp corresponding to the adjacent text reaches the preset value according to the converted text; When the time interval recorded in the recording time stamp corresponding to the adjacent text reaches the preset value, the segmentation module will record the time interval recorded in the corresponding recording time stamp reaching the preset value. Text boundaries divide text paragraphs.