阅读说明：本技术 一种会议纪要生成方法 (Conference summary generation method ) 是由 孔尧 于 2021-09-19 设计创作，主要内容包括：本发明提供一种会议纪要生成方法。具体可以包括：获取会议的音视频数据,识别并提取为文字信息；将提取的文字信息进行分类,分为指令文字信息和文本文字信息；对指令文字信息进行建模分类训练,识别指令文字信息并根据指令信息控制会议显示设备所呈现的内容；文本文字信息生成会议纪要。由于该会议纪要生成方法是将会议音视频中的识别的文字提取出指令文字信息和文本文字信息,通过指令文字信息控制会议设备,文本文字信息生成会议纪要或会议内容,该方法解决了开会时人为控制会议呈现内容、人为筛选并记录会议纪要、以及会议成员任务追踪时的各种问题。(The invention provides a conference summary generation method. The method specifically comprises the following steps: acquiring audio and video data of a conference, and identifying and extracting the audio and video data as character information; classifying the extracted character information into instruction character information and text character information; carrying out modeling classification training on the instruction character information, identifying the instruction character information and controlling the content presented by the conference display equipment according to the instruction information; the textual text information generates a conference summary. The conference summary generation method extracts the instruction character information and the text character information from the recognized characters in the conference audio and video, controls the conference equipment through the instruction character information, and generates the conference summary or the conference content from the text character information.)

1. A conference summary generation method, comprising:

s1: acquiring audio and video data of a conference, and identifying and extracting the audio and video data as character information;

s2: classifying the extracted character information into instruction character information and text character information;

s3: carrying out modeling classification training on the instruction character information, identifying the instruction character information and controlling the content presented by the conference display equipment according to the instruction information;

s4: the textual text information generates a conference summary.

2. The conference summary generation method according to claim 1, wherein the instruction text information in S3 includes:

the method comprises the following steps of equipment name information, action instruction information, switch input start-stop instructions, zoom-in and zoom-out instructions, screen switching instructions, skip instructions, equipment screen capturing instructions, insertion updating editing instructions and modification revising instructions.

3. The method for generating a conference summary according to claim 2, wherein after identifying the instruction text information and controlling the content presented by the conference display device according to the instruction text information in S3, the method further comprises:

when the conference display equipment is controlled to be the insertion updating editing instruction according to the instruction information, the insertion updating editing instruction is specifically to be inserted into screen capture information or text character information of another equipment, and after the equipment name model is identified according to the equipment name information, the screen capture information or the text character information of the equipment is inserted into a PPT (power point) presented by a conference summary.

4. The method for generating a conference summary according to claim 1, wherein when the text message generates a conference summary in step S4, the method further comprises:

and marking a conference speaker and distributing tasks for conference items according to the ID and name information of the conference members in the conference initiating APP.

5. The method for generating a conference summary according to claim 1, wherein when performing modeling classification training on the instruction text information, identifying the instruction text information, and controlling the content presented by the conference display device according to the instruction text information in S3, the method further comprises:

and carrying out multi-dimensional extraction modeling on the instruction character information, and carrying out instruction character classification training.

6. The method for generating a conference summary according to claim 5, wherein when performing multidimensional extraction modeling on the instruction text information, the method further comprises:

the multidimensional extraction modeling can be instructions of user voiceprints, voice pauses, voice continuous time and peripheral device names.

7. The method for generating a conference summary according to claim 1, wherein in the S3, the modeling classification training of the instruction text information, when identifying the instruction text information, further includes:

and modeling and classifying according to the instruction character information, and classifying the instruction character information similar to the natural language into a control instruction category.

8. The method of claim 7, wherein the modeling classification method further includes, but is not limited to:

k nearest neighbor classification algorithm, support vector machine and naive Bayes algorithm.

9. The method according to claim 1, wherein the step S4: after the text message generates the conference summary, the method further comprises:

and matching the conference members in the conference ID according to the text information in the conference summary to form conference tasks of the conference members.

10. The method of claim 9, wherein after forming the conference task for the conference member based on the conference members, the method further comprises:

and tracking the task progress of reminding the conference members until the conference members finish or end the conference task.

Technical Field

The invention relates to the technical field of communication, in particular to a conference summary generation method.

Background

With the rapid development of communication technology, various telephone conference and video conference devices are pushed to the market like a bamboo shoot in spring after rain, and among numerous conference devices, the conference display device, the conference call device and the conference control device are mainly used. The technician has more focus on the conference control equipment in order to better present and record the participants in the conference discussions or training content.

The method comprises the steps that human body actions and voiceprint information are obtained and used, in order to enable a speaker to present conference contents in a multi-direction mode, technical personnel obtain movement tracks and hand actions of the speaker through a camera arranged in a conference room, and action instructions are identified and converted into control instructions to control the presentation mode of the conference contents.

However, when meetings are carried out by using various meeting equipment, meeting preschool needs to be combined and presented manually, and the meeting preschool and task allocation are inevitably left or unclear.

Disclosure of Invention

The invention aims to provide a generation method of a conference summary, which is used for solving the problem that the conference summary and task allocation are missed or unclear because the conference summary needs to be manually combined and presented when various conference devices are used for meetings in the prior art.

In a first aspect, an embodiment of the present invention provides a method for generating a conference summary, which may include:

s1: acquiring audio and video data of a conference, and identifying and extracting the audio and video data as character information;

s2: classifying the extracted character information into instruction character information and text character information;

s3: carrying out modeling classification training on the instruction character information, identifying the instruction character information and controlling the content presented by the conference display equipment according to the instruction information;

s4: the textual text information generates a conference summary.

Further, the instruction text information in S3 may include:

the method comprises the following steps of equipment name information, action instruction information, switch input start-stop instructions, zoom-in and zoom-out instructions, screen switching instructions, skip instructions, equipment screen capturing instructions, insertion updating editing instructions and modification revising instructions.

Further, after the step S3 of recognizing the instruction text information and controlling the content presented by the conference display device according to the instruction information, the method may further include:

when the conference display equipment is controlled to be the insertion updating editing instruction according to the instruction information, the insertion updating editing instruction is specifically to be inserted into screen capture information or text character information of another equipment, and after the equipment name model is identified according to the equipment name information, the screen capture information or the text character information of the equipment is inserted into a PPT (power point) presented by a conference summary.

Further, when the text message generates a meeting summary in step S4, the method may further include:

and marking a conference speaker and distributing tasks for conference items according to the ID and name information of the conference members in the conference initiating APP.

Further, when performing modeling classification training on the instruction text information in S3, recognizing the instruction text information, and controlling the content presented by the conference display device according to the instruction information, the method may further include:

and carrying out multi-dimensional extraction modeling on the instruction character information, and carrying out instruction character classification training.

Further, when performing multidimensional extraction modeling on the instruction text information, the method may further include:

the multidimensional extraction modeling can be instructions of user voiceprints, voice pauses, voice continuous time and peripheral device names.

Further, the modeling and classification training of the instruction character information in S3 may further include, when identifying the instruction character information:

and modeling and classifying according to the instruction character information, and classifying the instruction character information similar to the natural language into a control instruction category.

Further, the modeling classification method may further include, but is not limited to:

k nearest neighbor classification algorithm, support vector machine and naive Bayes algorithm.

Further, the step S4: after the text message generates the conference summary, the method may further include:

and matching the conference members in the conference ID according to the text information in the conference summary to form conference tasks of the conference members.

Further, after forming a conference task for a conference member according to the conference member, the method may further include:

and tracking the task progress of reminding the conference members until the conference members finish or end the conference task.

The invention has the beneficial effects that:

in the conference summary generation method provided by the embodiment of the invention, audio and video data of a conference are acquired, identified and extracted as character information; classifying the extracted character information into instruction character information and text character information; carrying out modeling classification training on the instruction character information, identifying the instruction character information and controlling the content presented by the conference display equipment according to the instruction information; the textual text information generates a conference summary. The conference summary generation method extracts the instruction character information and the text character information from the recognized characters in the conference audio and video, controls the conference equipment through the instruction character information, and generates the conference summary or the conference content from the text character information.

Drawings

In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings needed to be used in the embodiments will be briefly described below, it should be understood that the following drawings only illustrate some embodiments of the present invention and therefore should not be considered as limiting the scope, and for those skilled in the art, other related drawings can be obtained according to the drawings without inventive efforts.

Fig. 1 is a schematic flow chart illustrating a conference summary generation method according to an embodiment of the present invention.

Detailed Description

In order to make the objects, technical solutions and advantages of the embodiments of the present invention clearer, the technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the drawings in the embodiments of the present invention, and it is obvious that the described embodiments are some, but not all, embodiments of the present invention. The components of embodiments of the present invention generally described and illustrated in the figures herein may be arranged and designed in a wide variety of different configurations.

Thus, the following detailed description of the embodiments of the present invention, presented in the figures, is not intended to limit the scope of the invention, as claimed, but is merely representative of selected embodiments of the invention. All other embodiments, which can be derived by a person skilled in the art from the embodiments given herein without making any creative effort, shall fall within the protection scope of the present invention.

The embodiment of the invention provides a conference summary generation method which can be applied to the fields of telephone conferences and video conferences and can realize the functions of automatic generation of conference summaries, automatic recording of conference contents, automatic tracking of conference tasks and the like.

Fig. 1 is a schematic flow chart illustrating a conference summary generation method according to an embodiment of the present invention.

As shown in fig. 1, an embodiment of the present invention provides a method for generating a conference summary, including:

s1: acquiring audio and video data of a conference, and identifying and extracting the audio and video data as character information;

s2: classifying the extracted character information into instruction character information and text character information;

s3: and carrying out modeling classification training on the instruction character information, identifying the instruction character information and controlling the content presented by the conference display equipment according to the instruction information.

S4: the textual text information generates a conference summary.

In one embodiment, when a user performs a multi-user teleconference or video conference, the conference devices accessed to the conference room include a television or a display, an electronic whiteboard, a camera, a sound extraction device such as a microphone, a sound player, a screen projector or a screen sharing device, and the like.

And initiating a multi-person telephone conference or a video conference by adopting third-party software, registering an ID (identity) and a nickname by each user, and accessing the conference according to the ID, the nickname and equipment information. And generating information such as conference subjects, conference IDs and the like of participants at the conference time and place, and synchronizing the conference information to the cloud, so that the conference members can modify, revise and the like conveniently.

S1: acquiring audio and video data of a conference, and identifying and extracting the audio and video data as character information;

specifically, audio and video data of the participants are extracted through conference room sound extraction equipment such as a microphone and the like and are converted into audio data, and the audio data are identified and extracted as character information.

S2: classifying the extracted character information into instruction character information and text character information;

specifically, the extracted character information is classified into instruction character information and text character information according to character content.

S3: carrying out modeling classification training on the instruction character information, identifying the instruction character information and controlling the content presented by the conference display equipment according to the instruction information;

specifically, modeling training is performed on the instruction character information, a specific control instruction is recognized, and the content presented by the conference display device is controlled according to the control instruction.

S4: the textual text information generates a conference summary.

Specifically, the text character information generates and stores meeting summary character information, and sends the meeting summary character information to a group corresponding to the meeting ID.

Optionally, the instruction text information in step S3 includes:

the method comprises the following steps of equipment name information, action instruction information, switch input start-stop instructions, zoom-in and zoom-out instructions, screen switching instructions, skip instructions, equipment screen capturing instructions, insertion updating editing instructions and modification revising instructions.

Optionally, after the step S3 identifies the instruction text information and controls the content presented by the conference display device according to the instruction information, the method further includes:

when the conference display equipment is controlled to be the insertion updating editing instruction according to the instruction information, the insertion updating editing instruction is specifically to be inserted into screen capture information or text character information of another equipment, and after the equipment name model is identified according to the equipment name information, the screen capture information or the text character information of the equipment is inserted into a PPT (power point) presented by a conference summary.

Optionally, when the text message generates a meeting summary in step S4, the method further includes:

and marking a conference speaker and distributing tasks for conference items according to the ID and name information of the conference members in the conference initiating APP.

Optionally, when the modeling classification training is performed on the instruction text information in step S3, and the instruction text information is recognized and the content presented by the conference display device is controlled according to the instruction information, the method further includes:

and carrying out multi-dimensional extraction modeling on the instruction character information, and carrying out instruction character classification training.

Optionally, when performing multi-dimensional extraction modeling on the instruction text information in step S3, the method may further include:

the multidimensional extraction modeling can be instructions of user voiceprints, voice pauses, voice continuous time and peripheral device names.

One specific embodiment: when multi-dimensional extraction modeling is carried out, the extracted features can be user voiceprint information, a voice pause rhythm, voice continuous time or voice duration time corresponding to the voiceprint information and instructions of peripheral equipment names. The training result may be: and the voiceprint information, the corresponding voice pause rhythm and the corresponding voice continuous identification instruction character information of a conference member of a certain user distribute peripheral equipment to execute the control instruction according to the name of the peripheral equipment in the conference room by using the identified instruction character information.

Optionally, the modeling and classification training of the instruction text information in step S3 may further include, when identifying the instruction text information:

and modeling and classifying according to the instruction character information, and classifying the instruction character information similar to the natural language into a control instruction category.

Optionally, the modeling classification method in step S3 may further include, but is not limited to:

a K nearest neighbor classification algorithm, a support vector machine, a naive Bayes and other classification algorithms.

In a specific embodiment, the instruction text information recognition modeling training adopts a K-Nearest Neighbor (KNN) classification algorithm to recognize a sample, such as a segment of audio, and most of K most similar samples (i.e., Nearest neighbors in a feature space) in a feature space belong to a certain class, and then the recognized sample is considered to belong to the class, and the content presented by the conference display device is controlled through a control instruction of the class.

For example: the participator speaks a 'next page' in the meeting place, because the recognized natural language has a pause, the participator judges the command text information, further recognizes the command text information, recognizes that the voice text information can be translated into different texts, such as 'next night', 'down page', 'next page' and other various combinations, classifies through a K nearest neighbor algorithm, recognizes different text combinations, matches a control command 'next page' in a command library, and executes a 'next page' action command in PPT or other application software in display equipment, thereby realizing the control of the content presented by the meeting display equipment according to the command information.

Namely: next night ≈ next page, execution: next night = next page instruction;

when the lower shift page is approximately equal to the next page, executing: move page down = next page instruction;

next page = next page, execution: next page = next page instruction.

Alternatively, step S4: after the text message generates the conference summary, the method may further include:

and matching the conference members in the conference ID according to the text information in the conference summary to form conference tasks of the conference members.

One specific embodiment: and matching the names or nickname information of the conference members in the conference group ID according to the text information recorded in the conference summary, and correspondingly forming the conference tasks of the conference members.

According to the method, the names or nicknames of the conference members in the conference group ID are matched according to the names or nicknames of the conference members mentioned during the conference, and the conference tasks or the conference contents are correspondingly recorded, so that the tasks or the contents related to the conference members are independently presented, the conference tasks of the conference members can be known only by looking at the conference summary or the conference contents related to the conference members, the time is saved, and the labor cost is saved.

Optionally, after forming a conference task of the conference member according to the conference member, the method may further include:

and tracking the task progress of reminding the conference members until the conference members finish or end the conference task.

One specific embodiment: and tracking the task completion progress according to the task of the conference member in the conference summary, reminding the member of completing the conference task in time, and knowing that the conference member completes or ends the task distributed by the conference.

The method can enable the meeting organizer or the meeting leader to know the progress condition of the meeting task in time and grasp the progress of the project task integrally.

Because the conference summary generation method extracts the instruction character information and the text character information from the recognized characters in the conference audio and video, the conference equipment is controlled by the instruction character information, and the text character information generates the conference summary or the conference content, the method greatly saves the presentation content of manually controlling the conference during the conference; matching the names of the conference members or the nicknames according to the conference ID, artificially screening and recording the conference summary; various problems during task tracking of conference members are solved.

The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention, and various modifications and changes may be made by those skilled in the art. Any modification, equivalent replacement, or improvement made within the spirit and principle of the present invention should be included in the protection scope of the present invention.