Meeting minutes recording method, recording device, electronic device and storage medium

Abstract

The present invention discloses a recording method, recording device, electronic device and storage medium for conference minutes. The recording method includes: continuously acquiring voice information; generating text information according to the voice information, the text information being the text form of the voice information; generating a hotspot in response to a hotspot recording signal, the hotspot corresponding to the currently acquired voice information; when it is determined that the acquisition of the voice information is completed, inputting the text information and the hotspot sequence into the conference minutes model, the hotspot sequence is formed by at least one hotspot, and the hotspot sequence is used to determine the part of the text information corresponding to the hotspot. The above-mentioned recording method and recording device convert voice information into text information, generate hotspots through hotspot recording signals, and combine text information and hotspots to mark the hotspot content in the text information, which is conducive to avoiding the hotspot content from being ignored and improving the quality of the conference minutes.

Description

Technical Field

[0001] The present invention relates to the field of multimedia communication, and in particular to a method, a recording device, an electronic device and a storage medium for recording meeting minutes. Background Art

[0002] With the popularity of modern office, conference software has become an important tool in office scenes. In conference scenes, there will be conversations containing a lot of content information. Text summarization technology can analyze the focus moments to help better understand the content of the meeting. In the existing technology, the focus moments are generally analyzed by obtaining the key information of the conversation, but there will be a problem of ignoring important moments and missing the focus content. Summary of the invention

[0003] The invention provides a method for recording meeting minutes, a recording device, an electronic device and a storage medium.

[0004] A method for recording meeting minutes provided by an embodiment of the present invention includes:

[0005] Continuously obtain voice information;

[0006] Generate text information according to the voice information, wherein the text information is a text form of the voice information;

[0007] generating a hotspot in response to a hotspot recording signal, the hotspot corresponding to the currently acquired voice information;

[0008] When it is determined that the voice information has been acquired, the text information and a hotspot sequence are input into a meeting minutes model, the hotspot sequence is formed by at least one hotspot, and the hotspot sequence is used to determine the portion of the text information corresponding to the hotspot.

[0009] The above recording method converts voice information into text information, generates hotspots through hotspot recording signals, and combines text information and hotspots to mark hotspot content in text information, which is helpful to avoid hotspot content being ignored and improve the quality of meeting minutes.

[0010] In some embodiments, the recording method comprises:

[0011] Get posture information of multiple participants;

[0012] When it is determined according to the posture information that the number of the participants who are looking at the speaker is greater than a preset number, the hotspot recording signal is generated.

[0013] In some embodiments, the recording method comprises:

[0014] In the case where it is determined that the hotspot recording control is operated, the hotspot recording signal is generated.

[0015] In some embodiments, the recording method comprises:

[0016] In the case of acquiring a hotspot recording gesture, generating the hotspot recording signal, wherein the hotspot recording gesture is determined by the posture information of the participant, or

[0017] In the case where a hotspot recording voice command is obtained, the hotspot recording signal is generated, and the hotspot recording voice command is determined by the voice information sent by the participant.

[0018] In some embodiments, inputting the text information and the hotspot sequence into a meeting minutes model includes:

[0019] Determining a plurality of sentence segments according to the text information;

[0020] The plurality of segments and the hotspot sequence are processed to adjust a weight of a first segment, the plurality of segments including the first segment, the first segment having the corresponding hotspot.

[0021] In some implementations, processing the multiple segments and the hotspot sequences includes:

[0022] Concatenating the multiple segments and the hotspot sequence to obtain multiple identified segments;

[0023] When the meeting minutes are generated according to the multiple identification segments and the meeting minutes model, the weight of the first segment is adjusted.

[0024] In some implementations, processing the multiple segments and the hotspot sequences includes:

[0025] determining the first sentence segment according to the plurality of sentence segments;

[0026] When the meeting minutes are generated according to the multiple segments and the meeting minutes model, the weight of the first segment is adjusted according to the hotspot sequence.

[0027] In some embodiments, inputting the text information and the hotspot sequence into a meeting minutes model includes:

[0028] Determining a plurality of sentence segments according to the text information;

[0029] The multiple sentence segments and the hotspot sequence are concatenated to obtain multiple sentence segments to be encoded, and the multiple sentence segments to be encoded are used to generate an environment text vector;

[0030] The environment text vector is decoded to obtain the output probability of each of the segments, the output probability of the first segment is greater than the output probability of the second segment, the multiple segments include the first segment and the second segment, the first segment has the corresponding hotspot, and the second segment does not have the corresponding hotspot.

[0031] An embodiment of the present invention provides a device for recording meeting minutes, wherein the device includes a control module, and the control module is used to:

[0032] Continuously obtain voice information;

[0033] Generate text information according to the voice information, wherein the text information is a text form of the voice information;

[0034] generating a hotspot in response to a hotspot recording signal, the hotspot corresponding to the currently acquired voice information;

[0035] When it is determined that the voice information has been acquired, the text information and a hotspot sequence are input into a meeting minutes model, the hotspot sequence is formed by at least one hotspot, and the hotspot sequence is used to determine the portion of the text information corresponding to the hotspot.

[0036] The above-mentioned recording device converts voice information into text information, generates hot spots through hot spot recording signals, and combines text information and hot spots to mark hot spot content in text information, which is helpful to avoid hot spot content being ignored and improve the quality of meeting minutes.

[0037] An electronic device provided by an embodiment of the present invention includes a memory and a processor, wherein the memory stores a computer program, and when the processor executes the computer program, the steps of the recording method described in any one of the above embodiments are implemented.

[0038] The above-mentioned electronic device converts voice information into text information, generates hotspots through hotspot recording signals, and combines text information and hotspots to mark hotspot content in text information, which is helpful to avoid hotspot content from being ignored and improve the quality of meeting minutes.

[0039] An embodiment of the present invention provides a computer-readable storage medium, on which a computer program is stored. When the computer program is executed by a processor, the steps of the recording method described in any one of the above embodiments are implemented.

[0040] The above-mentioned computer-readable storage medium converts voice information into text information, generates hotspots through hotspot recording signals, and combines text information and hotspots to mark hotspot content in text information, which is helpful to avoid hotspot content from being ignored and improve the quality of meeting minutes.

[0041] Additional aspects and advantages of the present invention will be given in part in the following description and in part will be obvious from the following description, or will be learned through practice of the present invention. BRIEF DESCRIPTION OF THE DRAWINGS

[0042] The above and / or additional aspects and advantages of the present invention will become apparent and easily understood from the description of the embodiments in conjunction with the following drawings, in which:

[0043] Figure 1 is a flow chart of a recording method according to an embodiment of the present invention;

[0044] Figure 2 It is a module schematic diagram of a recording device according to an embodiment of the present invention;

[0045] Figures 3 to 5 A schematic diagram of a data processing flow of an embodiment of the present invention;

[0046] Figure 6 FIG. 4 is another schematic diagram of a module of a recording device according to an embodiment of the present invention.

[0047] Description of main component symbols:

[0048] 100 - recording device; 110 - memory; 120 - processor; 130 - control module;

[0049] 210-text information; 211-sentence; 212-identified sentence; 213-sentence to be encoded; 214-environmental text vector; 220-hotspot; 221-hotspot sequence; 230-meeting minutes model; 231-meeting minutes; 240-input layer; 241-hidden layer; 242-output layer. DETAILED DESCRIPTION

[0050] The embodiments of the present invention are described in detail below, examples of which are shown in the accompanying drawings, wherein the same or similar reference numerals throughout represent the same or similar elements or elements having the same or similar functions. The embodiments described below with reference to the accompanying drawings are exemplary and are only used to explain the present invention, and cannot be understood as limiting the present invention.

[0051] See also Figure 1 , a method for recording meeting minutes 231 according to an embodiment of the present invention includes:

[0052] 01: Continuously obtain voice information;

[0053] 02: Generate text information 210 according to the voice information, where the text information 210 is a text form of the voice information;

[0054] 03: generating a hotspot 220 in response to the hotspot recording signal, where the hotspot 220 corresponds to the currently acquired voice information;

[0055] 04: When it is determined that the voice information has been acquired, the text information 210 and the hotspot sequence 221 are input into the meeting minutes model 230 , where the hotspot sequence 221 is formed by at least one hotspot 220 , and the hotspot sequence 221 is used to determine the portion of the text information 210 corresponding to the hotspot 220 .

[0056] The recording method of the embodiment of the present invention can be implemented by the recording device 100 of the meeting minutes 231 of the embodiment of the present invention. Figure 2 The recording device 100 includes a control module 130. The control module 130 is used to continuously obtain voice information. Text information 210 is generated according to the voice information, and the text information 210 is a text form of the voice information. A hotspot 220 is generated in response to a hotspot recording signal, and the hotspot 220 corresponds to the currently obtained voice information. When it is determined that the acquisition of the voice information is completed, the text information 210 and the hotspot sequence 221 are input into the meeting minutes model 230, and the hotspot sequence 221 is formed by at least one hotspot 220, and the hotspot sequence 221 is used to determine the part of the text information 210 corresponding to the hotspot 220.

[0057] The above-mentioned recording method and recording device 100 convert text information 210 through voice information, generate hotspot 220 through hotspot recording signal, and combine text information 210 and hotspot 220 to mark the hotspot content in text information 210, which is helpful to avoid hotspot content from being ignored and improve the quality of meeting minutes 231.

[0058] Specifically, see Figure 1 and Figure 2 In one embodiment, the conference scene may be an on-site meeting. The recording device 100 may record the voice information of the conference scene and obtain text information 210 through the voice information. At the same time, the recording device 100 responds to the hotspot recording signal of the conference scene. A hotspot 220 may be generated through the hotspot recording signal and a hotspot sequence 221 may be obtained. When the text information 210 and the hotspot sequence 221 are input into the meeting minutes model 230, the text information 210 corresponding to the hotspot sequence 221 may be obtained, and then the meeting minutes 231 integrated with the hotspot recording signal may be obtained. In this way, the missing hotspot content in the meeting minutes 231 may be avoided. In another embodiment, the conference scene may be a video conference.

[0059] In some embodiments, the recording method comprises:

[0060] Get posture information of multiple participants;

[0061] When it is determined according to the posture information that the number of participants who are looking at the speaker is greater than a preset number, a hotspot recording signal is generated.

[0062] The recording method of the embodiment of the present invention can be implemented by the recording device 100 of the meeting minutes 231 of the embodiment of the present invention. Figure 2 The control module 130 is used to obtain the posture information of multiple participants, and to generate a hot spot recording signal when it is determined according to the posture information that the number of participants who are looking at the speaker is greater than a preset number.

[0063] In this way, the hot spot recording signal can be obtained through posture judgment.

[0064] Specifically, in one embodiment, the conference device can obtain the image of the participant and obtain the posture information of the participant by image recognition, so as to determine whether the participant is currently looking at the speaker according to the posture information of the participant. If the number of participants who are looking at the speaker in the statistical conference scene is greater than a preset number, it can be determined that the speaker's current speech is a conference hotspot, so that a hotspot recording signal can be generated according to the voice information corresponding to the speaker's current speech, and the voice information in the conference scene can correspond to the hotspot content.

[0065] In some embodiments, the recording method comprises:

[0066] When it is determined that the hotspot recording control is operated, a hotspot recording signal is generated.

[0067] The recording method of the embodiment of the present invention can be implemented by the recording device 100 of the meeting minutes 231 of the embodiment of the present invention. Figure 2 The control module 130 is used to generate a hotspot recording signal when it is determined that the hotspot recording control is operated.

[0068] In this way, the hot spot recording signal can be obtained through active interaction.

[0069] Specifically, in one embodiment, the hotspot record control may be an interactive button. When some participants determine that the current voice information in the conference scene corresponds to the hotspot content, they may actively operate the interactive button, so that the control module 130 may respond to the participant's operation on the hotspot record control to generate a hotspot record signal.

[0070] In some embodiments, the recording method comprises:

[0071] When the hotspot recording posture is acquired, a hotspot recording signal is generated, and the hotspot recording posture can be determined by the posture information of the participant.

[0072] The recording method of the embodiment of the present invention can be implemented by the recording device 100 of the meeting minutes 231 of the embodiment of the present invention. Figure 2The control module 130 is used to generate a hotspot recording signal when a hotspot recording posture is acquired, and the hotspot recording posture is determined by the posture information of the participant.

[0073] In this way, a hot spot recording signal can be obtained by actively sending an interactive signal.

[0074] Specifically, in one embodiment, the recording device 100 can obtain the posture information of the participants in the conference scene. When some participants determine that the current voice information in the conference scene corresponds to the hot content, they can actively make a hot spot recording gesture. When the hot spot recording gesture is determined, the control module 130 generates a hot spot recording signal. In another embodiment, the recording device 100 can obtain the voice information of the participants in the conference scene. When some participants determine that the current voice information in the conference scene corresponds to the hot content, they can actively speak out the hot spot recording voice command. When the hot spot recording voice command is determined, the control module 130 generates a hot spot recording signal.

[0075] See also Figure 3 and Figure 4 In some embodiments, step 03 (inputting text information 210 and hotspot sequence 221 into meeting minutes model 230) includes:

[0076] Determine a plurality of segments 211 according to the text information 210;

[0077] The plurality of segments 211 and the hotspot sequence 221 are processed to adjust the weight of a first segment, the plurality of segments 211 including the first segment, and the first segment has a corresponding hotspot 220 .

[0078] The recording method of the embodiment of the present invention can be implemented by the recording device 100 of the meeting minutes 231 of the embodiment of the present invention. Figure 3 and Figure 4 The control module 130 is used to determine a plurality of segments 211 according to the text information 210 , process the plurality of segments 211 and the hotspot sequence 221 to adjust the weight of the first segment, the plurality of segments 211 including the first segment, and the first segment has a corresponding hotspot 220 .

[0079] In this way, hot content can be distinguished by weight.

[0080] Specifically, see Figure 3 and Figure 4In one embodiment, the control module 130 can obtain at least two segments 211 according to the text information 210. In combination with the hotspot sequence 221, the control module 130 can determine the segment 211 with the corresponding hotspot 220 as the first segment, and then adjust the weight of the first segment. Thus, the recording device can mark all the hotspot contents in the text information 210. In another embodiment, the meeting summary can be obtained by extracting the first segment in the text information 210.

[0081] See also Figure 3 In some implementations, the plurality of segments 211 and the hotspot sequence 221 are processed, including:

[0082] The multiple segments 211 and the hotspot sequence 221 are concatenated to obtain multiple identification segments 212;

[0083] When the meeting minutes 231 are generated according to the plurality of identification segments 212 and the meeting minutes model 230 , the weight of the first segment is adjusted.

[0084] The recording method of the embodiment of the present invention can be implemented by the recording device 100 of the meeting minutes 231 of the embodiment of the present invention. Figure 3 The control module 130 is used to concatenate the multiple segments 211 and the hotspot sequence 221 to obtain multiple identification segments 212. When the meeting minutes 231 are generated according to the multiple identification segments 212 and the meeting minutes model 230, the weight of the first segment is adjusted.

[0085] In this way, the input of the meeting minutes model 230 can be changed through the hotspot sequence 221 to achieve the adjustment of the meeting minutes 231 .

[0086] Specifically, see Figure 3 In one embodiment, the control module 130 can concatenate at least two segments 211 of the text information 210 with the hotspot sequence 221 at the input layer 240 to obtain at least two identification segments 212. The weight of the first segment with the corresponding hotspot 220 can be adjusted by combining the identification segment 212 with the meeting minutes model 230 at the hidden layer 241, and then the meeting minutes 231 with different weight tags can be obtained at the output layer 242.

[0087] See also Figure 4 In some implementations, the plurality of segments 211 and the hotspot sequence 221 are processed, including:

[0088] Determining a first segment according to the plurality of segments 211;

[0089] When the meeting minutes 231 are generated according to the plurality of segments 211 and the meeting minutes model 230 , the weight of the first segment is adjusted according to the hotspot sequence 221 .

[0090] The recording method of the embodiment of the present invention can be implemented by the recording device 100 of the meeting minutes 231 of the embodiment of the present invention. Figure 4 The control module 130 is configured to determine a first segment according to the plurality of segments 211 . When the meeting minutes 231 are generated according to the plurality of segments 211 and the meeting minutes model 230 , the weight of the first segment is adjusted according to the hotspot sequence 221 .

[0091] In this way, the weight of the segment 211 of the meeting minutes 231 may be adjusted according to the hot spot sequence 221 .

[0092] Specifically, see Figure 4 In one embodiment, the control module 130 may input at least two segments 211 in the input layer 240. The segment 211 and the meeting minutes model 230 may be combined to generate the meeting minutes 231 in the hidden layer 241. The meeting minutes 231 may be output after adjusting the weight of the first segment according to the hotspot sequence 221 in the output layer 242. Thus, the meeting minutes 231 with different weight marks may be obtained.

[0093] See also Figure 5 In some embodiments, step 03 (inputting text information 210 and hotspot sequence 221 into meeting minutes model 230) includes:

[0094] Determine a plurality of segments 211 according to the text information 210;

[0095] The plurality of sentence segments 211 and the hotspot sequence 221 are concatenated to obtain a plurality of to-be-encoded sentence segments 213, and the plurality of to-be-encoded sentence segments 213 are used to generate an environment text vector 214;

[0096] The environment text vector 214 is decoded to obtain the output probability of each sentence segment 211, the output probability of the first sentence segment is greater than the output probability of the second sentence segment, the multiple sentence segments 211 include the first sentence segment and the second sentence segment, the first sentence segment has a corresponding hotspot 220, and the second sentence segment does not have a corresponding hotspot 220.

[0097] The recording method of the embodiment of the present invention can be implemented by the recording device 100 of the meeting minutes 231 of the embodiment of the present invention. Figure 5The control module 130 is used to determine multiple segments 211 according to the text information 210. The multiple segments 211 and the hotspot sequence 221 are spliced ​​to obtain multiple segments to be encoded 213, and the multiple segments to be encoded 213 are used to generate the environment text vector 214. The environment text vector 214 is decoded to obtain the output probability of each segment 211, the output probability of the first segment is greater than the output probability of the second segment, the multiple segments 211 include the first segment and the second segment, the first segment has a corresponding hotspot 220, and the second segment does not have a corresponding hotspot 220.

[0098] In this way, hot content can be distinguished by outputting probability.

[0099] Specifically, see Figure 5 In one embodiment, the control module 130 can generate at least two segments 211 according to the text information 210. At least two to-be-encoded segments 213 can be obtained by splicing all the segments 211 with the hotspot sequence 221. The environment text vector 214 can be obtained by compiling all the to-be-encoded segments. The output probability of each segment 211 can be obtained by decoding the environment text vector 214 and outputting the meeting minutes 231. The control module 130 can determine that the segment 211 with the hotspot 220 is the first segment, and determine that the segment 211 without the hotspot 220 is the second segment. In another embodiment, the meeting summary can be obtained by extracting the first segment in the text information 210.

[0100] See also Figure 6 An electronic device according to an embodiment of the present invention includes a memory 110 and a processor 120. The memory 110 stores a computer program. When the processor 120 executes the computer program, the steps of the recording method according to any one of the above embodiments are implemented.

[0101] For example, when the computer program is executed, the following steps may be implemented:

[0102] 01: Continuously obtain voice information;

[0103] 02: Generate text information 210 according to the voice information, where the text information 210 is a text form of the voice information;

[0104] 03: generating a hotspot 220 in response to the hotspot recording signal, where the hotspot 220 corresponds to the currently acquired voice information;

[0105] 04: When it is determined that the voice information has been acquired, the text information 210 and the hotspot sequence 221 are input into the meeting minutes model 230 , where the hotspot sequence 221 is formed by at least one hotspot 220 , and the hotspot sequence 221 is used to determine the portion of the text information 210 corresponding to the hotspot 220 .

[0106] The above-mentioned electronic device converts text information 210 through voice information, generates hotspot 220 through hotspot recording signal, and combines text information 210 and hotspot 220 to mark the hotspot content in text information 210, which is helpful to avoid hotspot content from being ignored and improve the quality of meeting minutes 231.

[0107] A computer-readable storage medium according to an embodiment of the present invention stores a computer program, which, when executed by a processor, implements the steps of the recording method according to any one of the above embodiments.

[0108] For example, when the computer program is executed, the following steps may be implemented:

[0109] 01: Continuously obtain voice information;

[0110] 02: Generate text information 210 according to the voice information, where the text information 210 is a text form of the voice information;

[0111] 03: generating a hotspot 220 in response to the hotspot recording signal, where the hotspot 220 corresponds to the currently acquired voice information;

[0112] 04: When it is determined that the voice information has been acquired, the text information 210 and the hotspot sequence 221 are input into the meeting minutes model 230 , where the hotspot sequence 221 is formed by at least one hotspot 220 , and the hotspot sequence 221 is used to determine the portion of the text information 210 corresponding to the hotspot 220 .

[0113] The above-mentioned computer-readable storage medium converts text information 210 through voice information, generates hotspot 220 through hotspot recording signal, and combines text information 210 and hotspot 220 to mark the hotspot content in text information 210, which is helpful to avoid hotspot content from being ignored and improve the quality of meeting minutes 231.

[0114] The computer-readable storage medium may be provided in the recording device 100 or in other terminals. The recording device 100 may communicate with other terminals to obtain the corresponding program.

[0115] It can be understood that computer-readable storage media may include: any entity or device capable of carrying a computer program, recording media, USB flash drives, mobile hard disks, magnetic disks, optical disks, computer memory, read-only memory (ROM), random access memory (RAM), and software distribution media, etc. A computer program includes computer program code. The computer program code may be in source code form, object code form, executable file, or some intermediate form, etc. A computer-readable storage medium may include: any entity or device capable of carrying a computer program code, recording media, USB flash drives, mobile hard disks, magnetic disks, optical disks, computer memory, read-only memory (ROM), random access memory (RAM), and software distribution media.

[0116] In certain embodiments of the present invention, the module involved may be a single-chip microcomputer chip that integrates a processor, a memory, a communication module, etc. The processor may be a central processing unit (CPU), or other general-purpose processors, digital signal processors (DSP), application-specific integrated circuits (ASIC), field-programmable gate arrays (FPGA) or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc.

[0117] Any process or method description in a flowchart or otherwise described herein may be understood to represent a module, segment or portion of code that includes one or more executable instructions for implementing the steps of a specific logical function or process, and the scope of the preferred embodiments of the present invention includes alternative implementations in which functions may not be performed in the order shown or discussed, including performing functions in a substantially simultaneous manner or in the reverse order depending on the functions involved, which should be understood by those skilled in the art to which the embodiments of the present invention belong.

[0118] The logic and / or steps represented in the flowchart or otherwise described herein, for example, can be considered as an ordered list of executable instructions for implementing logical functions, and can be embodied in any computer-readable medium for use by an instruction execution system, apparatus or device (such as a computer-based system, a system including a processing module, or other system that can fetch instructions from an instruction execution system, apparatus or device and execute instructions), or used in combination with these instruction execution systems, apparatuses or devices.

[0119] Although embodiments of the present invention have been shown and described, it will be appreciated by those skilled in the art that various changes, modifications, substitutions and variations may be made to the embodiments without departing from the principles and spirit of the present invention, and that the scope of the present invention is defined by the claims and their equivalents.

Claims

1. A method for recording meeting minutes, characterized in that: include: Continuously obtain voice information; Generate text information according to the voice information, wherein the text information is a text form of the voice information; generating a hotspot in response to a hotspot recording signal, the hotspot corresponding to the currently acquired voice information; In the case where it is determined that the voice information is acquired, the text information and a hotspot sequence are input into a meeting minutes model, the hotspot sequence is formed by at least one hotspot, and the hotspot sequence is used to determine a portion of the text information corresponding to the hotspot; Inputting the text information and the hotspot sequence into a meeting minutes model includes: determining a plurality of segments according to the text information; processing the plurality of segments and the hotspot sequence to adjust the weight of a first segment, the plurality of segments include the first segment, the first segment has the corresponding hotspot, processing the plurality of segments and the hotspot sequence includes: concatenating the plurality of segments and the hotspot sequence to obtain a plurality of identified segments; in the case where the meeting minutes are generated according to the plurality of identified segments and the meeting minutes model, adjusting the weight of the first segment; or, inputting the text information and the hotspot sequence into a meeting minutes model includes: determining a plurality of segments according to the text information; concatenating the plurality of segments and the hotspot sequence to obtain a plurality of to-be-encoded segments, the plurality of to-be-encoded segments are used to generate an environment text vector; decoding the environment text vector to obtain an output probability of each of the segments, the output probability of the first segment is greater than the output probability of the second segment, the plurality of segments include the first segment and the second segment, the first segment has the corresponding hotspot, and the second segment does not have the corresponding hotspot.

2. The recording method according to claim 1, characterized in that The recording method comprises: Get posture information of multiple participants; When it is determined according to the posture information that the number of the participants who are looking at the speaker is greater than a preset number, the hotspot recording signal is generated.

3. The recording method according to claim 1, characterized in that The recording method comprises: In the case where it is determined that the hotspot recording control is operated, the hotspot recording signal is generated.

4. The recording method according to claim 1, characterized in that: The recording method comprises: In the case of acquiring a hotspot recording gesture, generating the hotspot recording signal, wherein the hotspot recording gesture is determined by the posture information of the participant, or In the case where a hotspot recording voice command is obtained, the hotspot recording signal is generated, and the hotspot recording voice command is determined by the voice information sent by the participant.

5. A device for recording meeting minutes, characterized in that: The recording device comprises a control module, and the control module is used for: Continuously obtain voice information; Generate text information according to the voice information, wherein the text information is a text form of the voice information; generating a hotspot in response to a hotspot recording signal, the hotspot corresponding to the currently acquired voice information; In the case where it is determined that the voice information is acquired, the text information and a hotspot sequence are input into a meeting minutes model, the hotspot sequence is formed by at least one hotspot, and the hotspot sequence is used to determine a portion of the text information corresponding to the hotspot; Inputting the text information and the hotspot sequence into a meeting minutes model includes: determining a plurality of segments according to the text information; processing the plurality of segments and the hotspot sequence to adjust the weight of a first segment, the plurality of segments include the first segment, the first segment has the corresponding hotspot, processing the plurality of segments and the hotspot sequence includes: concatenating the plurality of segments and the hotspot sequence to obtain a plurality of identified segments; in the case where the meeting minutes are generated according to the plurality of identified segments and the meeting minutes model, adjusting the weight of the first segment; or, inputting the text information and the hotspot sequence into a meeting minutes model includes: determining a plurality of segments according to the text information; concatenating the plurality of segments and the hotspot sequence to obtain a plurality of to-be-encoded segments, the plurality of to-be-encoded segments are used to generate an environment text vector; decoding the environment text vector to obtain an output probability of each of the segments, the output probability of the first segment is greater than the output probability of the second segment, the plurality of segments include the first segment and the second segment, the first segment has the corresponding hotspot, and the second segment does not have the corresponding hotspot.

6. An electronic device, characterized in that: The device comprises a memory and a processor, wherein the memory stores a computer program, and when the processor executes the computer program, the steps of the recording method according to any one of claims 1 to 4 are implemented.

7. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the steps of the recording method according to any one of claims 1 to 4 are implemented.

Images