Conference summary generation method and device and electronic equipment

Abstract

The invention relates to a conference summary generation method and device and electronic equipment, and the method comprises the steps: generating a first series of time stamps for a media stream of a conference in a conference process; monitoring an operation event using the document on the conference, and generating a second series of time stamps for the operation event; generating conference summary content based on the media stream with the first series of time stamps; in the conference summary content, an interaction control is configured for at least one content unit, and the interaction control is associated with at least one operation event and at least carries time stamp information of the associated operation event. Therefore, association among the conference summary, the media record and the document operation can be realized, so that the problems of low efficiency and difficulty in positioning of conference information backtracking in a traditional mode are thoroughly solved, and the backtracking efficiency and matching accuracy of conference related information are remarkably improved.

Description

Technical Field

[0001] This application relates to the field of computers, and more particularly to a method, apparatus and electronic device for generating meeting minutes. Background Technology

[0002] With the widespread adoption of remote collaboration and online work, online meetings have become a core scenario for corporate communication, solution review, and knowledge accumulation. In these meetings, sharing documents for joint review, discussion, and revision is a frequent and crucial activity. However, current technological solutions have significant shortcomings in achieving deep integration and efficient backtracking of meeting content and document operations, resulting in information silos.

[0003] First, traditional video conferencing software primarily relies on screen sharing to display documents. After the meeting, typically only the complete audio and video recordings are retained. If users want to review discussions on a specific section of the document, they must replay lengthy recordings and manually locate the relevant parts, a tedious and inefficient process. Second, while online collaborative documents support real-time editing by multiple users and version history viewing, these edit records are completely detached from the meeting context in which they were created. Users find it difficult to trace which meeting and discussion context a particular modification was made against, resulting in a loss of decision-making information. Summary of the Invention

[0004] This application provides a method, apparatus, and electronic device for generating meeting minutes to solve the technical problem in the prior art where meeting minutes are separated from the operation records of documents used in the meeting, making it difficult to trace and locate meeting information.

[0005] Firstly, this application provides a method for generating meeting minutes, the method comprising: During the meeting, a first series of timestamps are generated for the media stream of the meeting; operation events using documents in the meeting are listened for, and a second series of timestamps are generated for the operation events; Based on the media stream with the first series of time stamps, generate meeting minutes content; In the meeting minutes, at least one content unit is configured with an interactive control, which is associated with at least one operation event and carries at least the time stamp information of the associated operation event.

[0006] In one possible implementation, the method further includes: In response to a trigger operation on the interactive control, a first linkage backtracking operation is executed; The first linkage backtracking operation includes: causing the media stream to start displaying at the corresponding time point according to the time stamp information carried by the interactive control, and controlling the document to present the corresponding document state according to the operation events associated with the interactive control.

[0007] In one possible implementation, the method further includes: In response to detecting an operation event of a document used in the meeting, a document snapshot corresponding to the operation event is obtained, the document snapshot containing at least the document state information of the document at the time the operation event occurred; The step of controlling the document to present a corresponding document state based on the operation events associated with the interactive control includes: Based on the operation events associated with the interactive control, obtain the corresponding document snapshot; Based on the obtained document snapshot, control the document to present the corresponding document state.

[0008] In one possible implementation, controlling the document to present a corresponding document state includes: Locate and display the document at the corresponding document position; And / or, presenting document content change events triggered by operation events corresponding to the document snapshot.

[0009] In one possible implementation, the method further includes: Based on the monitored operation events, generate the document's document editing history; In response to viewing a target operation event in the document editing history, a second linked backtracking operation is executed, the second linked backtracking operation including at least one of the following: Based on the time stamp information of the target operation event, locate and present the content unit associated with the target operation event in the meeting minutes content; Based on the time stamp information of the target operation event, the media stream is displayed at the corresponding time point.

[0010] In one possible implementation, the method further includes: While displaying the media stream, the content unit corresponding to the current display period is highlighted in the display interface of the meeting minutes content.

[0011] In one possible implementation, the associated operation events for the content units in the meeting minutes are determined in the following manner: Based on the media stream with the first series of time stamps, determine the time window corresponding to the content unit in the meeting minutes content; At least one operation event whose time stamp falls within the time window is identified as an operation event associated with the content unit.

[0012] In one possible implementation, the operation event that causes the timestamp to fall within the time window, determined as an operation event associated with the content unit, includes: Operation events whose timestamps fall within the time window are identified as candidate operation events. Determine the semantic relevance between the candidate operation events and the content unit, and based on the semantic relevance, determine the operation events associated with the content unit from the candidate operation events.

[0013] In one possible implementation, the method further includes: During the meeting, based on the audio in the media stream, the speaker is identified using voiceprint recognition technology, and the start and end time markers of the speaker's corresponding speaking segment are determined. Based on the start and end time markers, the speaker identifier is associated with the corresponding content unit in the meeting minutes.

[0014] In one possible implementation, the method further includes: In response to the triggering operation of the interactive control, a first linkage backtracking operation is performed, and the speaker's identifier associated with the content unit corresponding to the interactive control is displayed in the display interface of the meeting minutes content.

[0015] In one possible implementation, configuring interactive controls for at least one content unit in the meeting minutes includes: In response to determining that the same content unit in the meeting minutes is associated with operation events for multiple different documents, multiple interactive controls are configured for the content unit, each of which is associated with a different document and a corresponding operation event.

[0016] Secondly, this application provides a meeting minutes generation apparatus, the apparatus comprising: The first time stamp module is used to generate a first series of time stamps for the media stream of the conference during the conference process; A document operation monitoring module is used to monitor operation events of documents used in the meeting and generate a second series of timestamps for the operation events. The meeting minutes generation module is used to generate meeting minutes content based on the media stream with the first series of time stamps; The document operation association module is used to configure interactive controls for at least one content unit in the meeting minutes content. The interactive controls are associated with at least one operation event and at least carry the time stamp information of the associated operation event.

[0017] In one possible implementation, the device further includes: The first linkage backtracking module is used to execute the first linkage backtracking operation in response to the trigger operation of the interactive control; The first linkage backtracking operation includes: causing the media stream to start displaying at the corresponding time point according to the time stamp information carried by the interactive control, and controlling the document to present the corresponding document state according to the operation events associated with the interactive control.

[0018] In one possible implementation, the device further includes: A document snapshot generation module is used to obtain a document snapshot corresponding to the operation event in response to listening to an operation event of a document used in the meeting. The document snapshot at least includes the document state information of the document at the time the operation event occurs. The first linkage backtracking module controls the document to present a corresponding document state based on the operation events associated with the interactive control, including: Based on the operation events associated with the interactive control, obtain the corresponding document snapshot; Based on the obtained document snapshot, control the document to present the corresponding document state.

[0019] In one possible implementation, the first linkage backtracking module controls the document to present a corresponding document state, including: Locate and display the document at the corresponding document position; And / or, presenting document content change events triggered by operation events corresponding to the document snapshot.

[0020] In one possible implementation, the device further includes: The editing history generation module is used to generate the document editing history of the document based on the monitored operation events; The second linkage backtracking module is used to execute a second linkage backtracking operation in response to a viewing operation of the target operation event in the document editing history. The second linkage backtracking operation includes at least one of the following: Based on the time stamp information of the target operation event, locate and present the content unit associated with the target operation event in the meeting minutes content; Based on the time stamp information of the target operation event, the media stream is displayed at the corresponding time point.

[0021] In one possible implementation, the device further includes: The highlighting module is used to highlight the content unit corresponding to the current display period in the display interface of the meeting minutes while displaying the media stream.

[0022] In one possible implementation, the document operation association module is used to determine associated operation events for content units in the meeting minutes content in the following manner: Based on the media stream with the first series of time stamps, determine the time window corresponding to the content unit in the meeting minutes content; At least one operation event whose time stamp falls within the time window is identified as an operation event associated with the content unit.

[0023] In one possible implementation, the document operation association module determines at least one operation event whose timestamp falls within the time window as an operation event associated with the content unit, including: Operation events whose timestamps fall within the time window are identified as candidate operation events. Determine the semantic relevance between the candidate operation events and the content unit, and based on the semantic relevance, determine the operation events associated with the content unit from the candidate operation events.

[0024] In one possible implementation, the device further includes: The speaker association module is used to identify speakers based on the audio in the media stream using voiceprint recognition technology during the meeting process, and to determine the start and end time markers of the speaking segments corresponding to the speaker. Based on the start and end time markers, the speaker identifier is associated with the corresponding content unit in the meeting minutes.

[0025] In one possible implementation, the device further includes The speaker display module is used to respond to the trigger operation of the interactive control, perform a first linkage backtracking operation, and display the speaker identifier associated with the content unit corresponding to the interactive control in the display interface of the meeting minutes content.

[0026] In one possible implementation, the document operation association module configures interactive controls for at least one content unit in the meeting minutes content, including: In response to determining that the same content unit in the meeting minutes is associated with operation events for multiple different documents, multiple interactive controls are configured for the content unit, each of which is associated with a different document and a corresponding operation event.

[0027] Thirdly, this application provides an electronic device, including a processor and a memory, wherein the processor is configured to execute a meeting minutes generation program stored in the memory to implement the meeting minutes generation method described in any one of the first aspects.

[0028] Fourthly, this application provides a storage medium storing one or more programs that can be executed by one or more processors to implement the meeting minutes generation method described in any one aspect.

[0029] Compared with the prior art, the technical solution provided in this application has the following advantages: The method provided in this application generates a first series of time stamps for the media stream of the meeting during the meeting process, listens to the operation events of the documents used in the meeting, and generates a second series of time stamps for the operation events. This realizes the time correlation between the two heterogeneous data types, the meeting media stream and the document operation events, fundamentally solving the core pain point of the prior art where meeting minutes, media recordings, and document operations are separated and the time dimensions cannot correspond, laying a precise time foundation for the correlation of the three; based on the media stream with the first series of time stamps, meeting minutes are generated. Content: In the meeting minutes, at least one content unit is configured with an interactive control. The interactive control is associated with at least one operation event and carries at least the time stamp information of the associated operation event. This establishes a direct link between the meeting minutes, document operation events, and media streams. Unlike traditional technologies, users do not need to manually browse through massive amounts of documents or fast-forward and rewind video recordings to match relevant information. They can quickly link to the corresponding document operation records and media stream segments simply by triggering the interactive control. This completely solves the problems of inefficient meeting information retrieval and difficult location in the traditional model, and significantly improves the efficiency and accuracy of meeting-related information retrieval and matching. Attached Figure Description

[0030] The accompanying drawings, which are incorporated in and form part of this specification, illustrate embodiments consistent with this application and, together with the description, serve to explain the principles of this application.

[0031] To more clearly illustrate the technical solutions in the embodiments of this application or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, for those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0032] One or more embodiments are illustrated by way of example with reference numerals in the accompanying drawings. These illustrations do not constitute a limitation on the embodiments. Elements with the same reference numerals in the drawings are denoted as similar elements. Unless otherwise stated, the figures in the drawings are not to be limited by scale.

[0033] Figure 1 A flowchart illustrating an embodiment of a meeting minutes generation method provided in this application; Figure 2A flowchart illustrating an embodiment of another meeting minutes generation method provided in this application; Figure 3 A flowchart illustrating an embodiment of another meeting minutes generation method provided in this application; Figure 4 A block diagram illustrating an embodiment of a meeting minutes generation device provided in this application; Figure 5 This is a schematic diagram of the structure of an electronic device provided in an embodiment of this application. Detailed Implementation

[0034] To make the objectives, technical solutions, and advantages of the embodiments of this application clearer, the technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this application, not all embodiments. Based on the embodiments of this application, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this application.

[0035] The following disclosure provides numerous different embodiments or examples for implementing various structures of this application. To simplify the disclosure, specific examples of components and arrangements are described below. These are merely examples and are not intended to limit the scope of this application. Furthermore, reference numerals and / or letters may be repeated in different examples. Such repetition is for simplification and clarity and does not in itself indicate a relationship between the various embodiments and / or arrangements discussed.

[0036] To address the technical problem in existing technologies where meeting minutes and operation records of documents used in meetings are separated, making it difficult to trace and locate meeting information, this application provides a meeting minutes generation method, apparatus, and electronic device that can link meeting minutes, media records, and document operations. This completely solves the problems of inefficient meeting information retrieval and difficulty in locating information in the traditional model, and significantly improves the retrieval efficiency and matching accuracy of meeting-related information.

[0037] Figure 1This is a flowchart illustrating an embodiment of a meeting minutes generation method provided in this application. In one embodiment, the executing entity of this application embodiment can be a device with data processing and network communication capabilities. The device can be a hardware entity or a software functional module deployed on a hardware entity. When the device is a hardware entity, it can be various electronic devices with displays, including but not limited to smartphones, tablets, laptops, desktop computers, servers, etc. When the device is a software functional module, it can be installed and run on the aforementioned electronic devices, providing corresponding network services through client applications or server applications. Specifically, if the device installs and runs a client application, it can act as a user device in network communication; if the device installs and runs a server application, it can act as a server in network communication, providing corresponding services. In the application scenarios involved in this application, the services include, but are not limited to, online meeting services. Figure 1 As shown, the method includes the following steps: Step 101: During the meeting process, generate a first series of timestamps for the media stream of the meeting, and listen for operation events of documents used in the meeting, and generate a second series of timestamps for the operation events.

[0038] An exemplary application scenario of this application is an online meeting scenario, that is, a remote meeting format that connects participants distributed in different geographical locations through a network. It covers scenarios that require simultaneous operation of documents and recording of the meeting process, such as internal enterprise document review, cross-regional solution discussion, and multi-terminal collaborative review.

[0039] During the meeting, audio, video, and screen sharing content will be continuously recorded to form a continuous media stream. For example, if the executing entity of this embodiment is a user device (such as a desktop computer), the user device can directly and continuously record audio and video and capture screen sharing content through its integrated audio and video recording and screen capture modules to form a continuous media stream. If the executing entity of this embodiment is a server, the user device can continuously record audio, video, and screen sharing content to form a continuous media stream and upload the generated media stream to the server in real time. Subsequently, the server will lead the execution of the technical solution provided in this embodiment. Furthermore, in the scenario where the executing entity of this embodiment is a server, the subsequent steps may also involve interaction and collaboration between the user device and the server (such as the server pushing related data to the user device, the user device responding to an operation and then initiating a request to the server), thereby implementing the technical solution provided in this embodiment. Those skilled in the art can logically deduce this based on actual application scenarios and deployment requirements; this embodiment will not elaborate further.

[0040] The first series of timestamps is an ordered and continuous sequence of timestamps injected into the media stream of the conference, where each timestamp corresponds to a moment in a frame or segment of the media stream. These timestamps are generated based on a unified clock source (e.g., with the system time at the start of the conference as the zero point of the timeline) and together form the reference timeline for aligning all subsequent document operation events.

[0041] Documents used in a meeting include both shared documents and documents that participants can view independently on their local devices. For example, multiple documents can be uploaded to the meeting space with corresponding permissions configured, allowing participants to view these documents according to their needs and permissions. When participants view documents locally on their clients, the documents can be displayed in view or interface format on their clients, but are not visible to other participants. These documents still fall under the category of documents used in the meeting.

[0042] Document operation events refer to various interactive behaviors performed by participants on documents used in the meeting. These include not only content editing operations (insertion / deletion / modification, etc.), but also operations such as opening / closing the document, switching pages, scrolling, selecting areas, adding annotations or comments, switching view modes, and adjusting zoom levels. Furthermore, in one embodiment, document operation events also include user-initiated manual marking events. These events are triggered by the user to indicate and record the current document's specific state. For example, when a participant clicks the "Mark" button or issues a voice command such as "Mark this page," the recording of this marking event is triggered, and the current document's specific state (such as page number, scroll position, and content presentation format) is simultaneously retained as the core content of the marking event.

[0043] For example, document operation event listening can be implemented in several ways: First, using a browser plugin, developing a corresponding browser plugin and injecting JavaScript code into the target page such as the document; second, using DOM (Document Object Model) event listening, leveraging MutationObserver to listen for DOM changes in the document content area, thereby capturing various editing operations; third, implementing user interaction event listening, using event listeners, such as addEventListener, to listen for user operation events such as scroll, click, and keydown, to capture page switching, scrolling, and other behaviors; fourth, using API (Application Programming Interface) interfaces, calling the API interfaces provided by document platforms such as Kingsoft Docs to subscribe to push notifications of document operation events; fifth, establishing long-lived connections using communication protocols, such as WebSocket long-lived connections, to maintain continuous communication with the document service and receive relevant notifications of document operation events in real time. Besides the above online listening implementation methods, there are also corresponding alternative implementation schemes, such as video recording combined with OCR (Optical Character Recognition): by recording a complete screen video, document operation-related information can be extracted after the meeting using OCR and image recognition technologies. The advantage of this approach is that it does not require integration with a document API and is applicable to various document types. This application does not limit the specific method used to implement document operation event listening in its embodiments.

[0044] The second series of time stamps are generated in real-time for each monitored document operation event. The sequence of these time stamps is based on the same time reference as the "first series of time stamps," ensuring precise time alignment between discrete document operations and continuous media streams. Furthermore, for network transmission events such as remote participant operations, network latency can be recorded and compensated for in the time stamps, further guaranteeing time alignment accuracy.

[0045] As described above, the following two core actions will be executed in parallel during the meeting: First, the media stream of the meeting will be continuously recorded, and the first series of timestamps will be generated simultaneously for the media stream of the meeting, fully recording the content of the meeting discussion and the corresponding timeline; Second, all operation events for the documents used in the meeting will be monitored and captured in real time. Each operation event will immediately generate an operation event record with timestamps, accurately recording every moment of change in the document status and the specific content of the change.

[0046] For example, the operation log could be presented as [00:54:32|Q3 Sales Report.pptx|Page 5|Editing Operation: Change the title "Sales Data" to "Q3 Regional Sales Data Comparison"|Zhang San]. This means that at 00:54:32 (based on the relative time stamp of the meeting start time T0), participant "Zhang San" performed an editing operation on the document named "Q3 Sales Report.pptx". This editing operation occurred on page 5 of the document, and the specific editing content was "change the title 'Sales Data' to 'Q3 Regional Sales Data Comparison'". Those skilled in the art will understand that this is merely an exemplary presentation of the operation log. In practical applications, the operation log can be designed to include more information as needed, such as document snapshot identifiers before and after the operation, document view modes (such as "presentation mode"), etc. This application embodiment does not limit this.

[0047] By attaching precisely comparable time stamps originating from the same clock source to both continuous media streams and discrete operation events, step 101 establishes a temporal link between the conference media streams and document operation events. This lays a precise time alignment foundation for subsequent intelligent semantic association and cross-media collaborative backtracking operations.

[0048] Step 102: Generate meeting minutes based on the media stream with the first series of time stamps.

[0049] The meeting minutes refer to the structured record generated after performing speech recognition, natural language processing, and multi-format content integration on the media stream with the first series of time stamps. For example, the meeting minutes are not simply a full transcript of the audio, but a comprehensive record that has been systematically summarized and optimized, integrating various presentation formats such as text, images, and lists. It may include elements such as speech segments, speaker annotations, topic extraction, key information extraction, key document snapshots, and to-do lists.

[0050] In this embodiment, regardless of the presentation format, the meeting minutes inherit the temporal attributes of the source audio and video segments. That is, each text paragraph, each configuration image, and each list item corresponds to a specific time point or time interval. For example, if the audio segment "10 minutes 15 seconds to 10 minutes 30 seconds" is identified as saying "Let's look at the chart on page five," then the generated minutes text "Let's look at the chart on page five" is associated with the time interval [00:10:15-00:10:30].

[0051] In one embodiment, the specific implementation of generating meeting minutes based on a media stream with a first series of time stamps includes: First, using speech recognition technology, the audio stream with the first series of time stamps is accurately converted into text data, ensuring consistency between the speech content and the text transcription; during the speech-to-text conversion and subsequent content integration process, the first series of time stamps in the original audio stream will be accurately mapped to all forms of minutes content; then, through natural language processing technology and content structuring algorithms, the transcribed text and associated data are optimized and integrated in multiple forms, where the text portion can be segmented according to the speaking logic, the speaker's identity can be identified, and the discussion topics and key viewpoints can be extracted to form a clear and organized text record, while embedding document snapshot images of key operation nodes during the meeting (such as page switching, important annotations, screenshots of the edited document status, etc.) to intuitively restore the document presentation scene during the discussion. In addition, scattered key conclusions, to-do items, points of contention, action plans, etc., can be organized into ordered or unordered lists to improve information readability. Finally, all forms of minutes content are arranged in an orderly manner according to the timeline, and text, images, and lists within the same time interval are interconnected to form a complete record of meeting scene segments.

[0052] Step 103: In the meeting minutes, configure an interactive control for at least one content unit. The interactive control is associated with at least one operation event and carries at least the time stamp information of the associated operation event.

[0053] Among them, the content unit refers to the smallest information unit with independent time attributes in the meeting minutes content generated in step 102, including but not limited to a single paragraph of text, a single embedded image, a single list item, etc. Each content unit is associated with a specific time point (such as 00:15:20) or time interval (such as [00:10:15-00:10:30]).

[0054] Interactive controls are operable interactive elements (which can take the form of time anchors, icons, text links, etc.) configured within content units. Their function is to serve as a link between meeting minutes content and document operation events, allowing users to invoke the corresponding document status and meeting audio / video through clicks, triggers, and other operations. For example, an interactive control is associated with at least one operation event and carries the time stamp information of the associated operation event (e.g., [00:10:35]). The configured interactive control has interactive response attributes. When a user triggers the control (e.g., by clicking a time anchor), it can read the associated information it carries, providing data support for subsequent linked operations such as initiating audio playback and document location.

[0055] Therefore, in step 103, configuring interactive controls for at least one content unit in the meeting minutes aims to bind the "minute content with time attributes" with the "document operation events with time stamps" through interactive controls, breaking down the information barriers between the meeting minutes, the documents used in the meeting, and the meeting audio and video streams, allowing users to achieve the "click to rewind" function through interactive controls, and laying the interactive foundation for subsequent cross-media collaboration.

[0056] Furthermore, as mentioned earlier, the documents used in the meeting can themselves have access permission configurations. Correspondingly, the interactive controls configured for content units in the meeting minutes can inherit the permission attributes of their associated documents. That is, for users who do not have access to the corresponding document, the interactive controls associated with that document will either be invisible or ineffective when triggered by the user (unable to recall the document state and media stream rewind), thereby ensuring the consistency of document permission control, protecting the security of document data access, and preventing unauthorized access and data leakage.

[0057] In one embodiment, the associated operation events for content units in the meeting minutes are determined by: determining the time window corresponding to the content unit in the meeting minutes based on the media stream with a first series of time stamps; and determining at least one operation event whose time stamp falls within the time window as the operation event associated with the content unit.

[0058] The core logic of this embodiment relies on the timeline alignment characteristics between media streams, meeting minutes content, and operation events to bind operation events to content units through time window filtering. Specifically, firstly, since meeting minutes content units (such as single discussion text or single list items) are generated based on media streams with a first series of time stamps, each content unit inherits the time attributes of the source media stream and corresponds to specific time stamp information, such as the time interval [T_start, T_end]. This time interval is the "time window" of the content unit. For example, if a content unit originates from a meeting audio segment from 10 minutes 15 seconds to 10 minutes 30 seconds, its corresponding time window is [10:15, 10:30], precisely corresponding to the time range in which the discussion content occurred during the meeting. Subsequently, all document operation events recorded in step 101 (all carrying a second series of time stamps with the same origin as the first series of time stamps) are traversed, and at least one operation event whose second series of time stamps fall within the aforementioned time window is directly identified as the operation event associated with the content unit. The core basis of this matching logic is the "strong temporal correlation between meeting discussions and document operations": In a meeting scenario, the discussion content of the participants usually revolves around the document they are currently operating on or viewing. The document operations within the time window (such as page switching, editing, annotation, etc.) are directly related to the discussion topic of the corresponding content unit. The design of the first series of time markers and the second series of time markers based on the same clock source and with the meeting start time T0 as the reference ensures the accuracy of time matching and avoids correlation deviations caused by time errors.

[0059] For example, if the time window of a content unit is [10:15, 10:30] and the discussion content is "analyzing the regional sales chart on page 5", then the document operation events with the second series of timestamps within that time window (such as "page switched to page 5" at 10:20 and "chart highlighted" at 10:25) will be filtered out, and these operation events will be identified as the associated objects of that content unit.

[0060] Further, in one embodiment, determining at least one operation event whose time marker falls within a time window as an operation event associated with a content unit includes: determining the operation event whose time marker falls within the time window as a candidate operation event; determining the semantic correlation between the candidate operation event and the content unit; and determining the operation event associated with the content unit from the candidate operation events based on the semantic correlation.

[0061] This embodiment, building upon the previous one, adds semantic relevance verification to the time-based filtering to eliminate operation events within the time window that are irrelevant to the discussion topic of the content unit and ensure the accuracy of the association results. The specific implementation logic is as follows: First, all document operation events whose timestamps fall within the time window of the content unit are uniformly identified as "candidate operation events." This step retains all time-related operation events, avoiding the omission of potential associated objects. Subsequently, core semantic information is extracted from the meeting minutes content unit, including discussion keywords (such as "page 5," "chart," "modify title," etc.) and discussion topics (such as document content analysis, format adjustment, etc.). At the same time, key descriptive information for each candidate operation event is extracted (such as operation type "page switching," operation object "page 5," operation content "annotate chart," etc.). Then, the semantic similarity between the two is calculated using natural language processing technology. Finally, a semantic relevance threshold is set (which can be flexibly adjusted according to the actual scenario), and candidate operation events with semantic similarity exceeding the threshold are identified as the operation events ultimately associated with the content unit. If there are multiple candidate events that exceed the threshold (such as "page switched to page 5" and "annotate the chart on page 5" within the same time window, both of which are highly relevant to the discussion topic), then all of them will be treated as related events, and multiple interactive controls can be configured accordingly.

[0062] This progressive design ensures the efficiency of association through time filtering and the accuracy of association through semantic verification. It fits the actual scenario where "meeting discussions and document operations are related in both time sequence and topic", avoiding invalid associations that may occur when relying solely on time matching, such as "time-related but topic-irrelevant" associations. It provides more user-friendly association data support for subsequent interactive linkages.

[0063] The technical solution provided in this application, by generating a first series of time stamps for the media stream during the meeting process, listening to the operation events of documents used in the meeting, and generating a second series of time stamps for the operation events, realizes the time association between the two heterogeneous data types of meeting media stream and document operation events. This fundamentally solves the core pain point of the prior art where meeting minutes, media records, and document operations are mutually isolated and the time dimensions cannot correspond, laying a precise time foundation for the association of the three. Based on the media stream with the first series of time stamps, meeting minutes content is generated. In the meeting minutes content, at least one content unit is configured with an interactive control. The interactive control is associated with at least one operation event and carries at least the time stamp information of the associated operation event. This realizes a direct association channel between meeting minutes, document operation events, and media stream. Users do not need to manually browse through massive documents or fast-forward and rewind video recordings to match relevant information as in traditional technology. They can quickly associate the corresponding document operation records and media stream segments simply by triggering the interactive control. This completely solves the problems of inefficient meeting information backtracking and difficult location in the traditional mode, and significantly improves the backtracking efficiency and matching accuracy of meeting-related information.

[0064] Figure 2 A flowchart illustrating an embodiment of another meeting minutes generation method provided in this application. Figure 2 The process shown is in Figure 1 Based on the illustrated process, the following steps are included: Step 201: During the meeting process, generate a first series of timestamps for the media stream of the meeting, and listen for operation events of documents used in the meeting, and generate a second series of timestamps for the operation events.

[0065] Step 202: Generate meeting minutes based on the media stream with the first series of time stamps.

[0066] Step 203: In the meeting minutes, configure an interactive control for at least one content unit. The interactive control is associated with at least one operation event and carries at least the time stamp information of the associated operation event.

[0067] For detailed explanations of steps 201 to 203, please refer to the above. Figure 1 The relevant descriptions in the process shown will not be repeated here.

[0068] Step 204: In response to the trigger operation of the interactive control, execute the first linkage backtracking operation.

[0069] The first linked backtracking operation includes: based on the time stamp information carried by the interactive control, causing the media stream to start displaying at the corresponding time point, and controlling the document to present the corresponding document state based on the operation events associated with the interactive control. For example, controlling the document to present the corresponding document state includes: locating and presenting the document at the corresponding document position; and / or, presenting the document content change event triggered by the operation event corresponding to the document snapshot.

[0070] As can be seen from the above description, in step 204, the execution of the first linkage backtracking operation in response to the triggering operation of the interactive control is the core execution link to realize the three-dimensional linkage of "meeting minutes content - document - media stream". It can accurately restore the discussion context and / or document status at the time of the meeting through the triggering operation of the interactive control configured in the user's meeting minutes content, and completely solve the problems of information fragmentation and backtracking difficulties in traditional technologies.

[0071] Specifically, when a user views meeting minutes and performs a trigger operation on an interactive control embedded in a content unit (such as clicking a time anchor or document association icon), the execution entity in this embodiment will immediately respond and initiate a linked backtracking process. For example: using the time stamp information carried by the interactive control as an index, the system locates the corresponding position in the meeting media stream based on this time stamp information and begins displaying the media stream from that located position. If the media stream contains audio, it automatically jumps to the corresponding time to start audio playback, restoring the context of the meeting discussion, speakers' dialogue, etc.; if the media stream contains screen-sharing video, it synchronously plays the video footage at the corresponding time, forming a dual scene restoration of visual and auditory elements with the audio and document status. Furthermore, during playback, the context of the minutes before and after the playback time point can also be displayed in conjunction with the playback time point. For example, the content unit corresponding to the current display period (such as the corresponding speaking paragraph) can be highlighted in the meeting minutes content display interface, helping users quickly connect to the discussion logic.

[0072] Meanwhile, based on the operation events associated with the interactive controls, the operation event data recorded in step 101 is invoked to control the document presentation to be consistent with the state at the time of the meeting. This includes two presentation formats (which can be executed individually or simultaneously): First, document positioning and basic state restoration: The target document associated with the interactive controls (such as "Q3 Sales Report.pptx") is automatically opened, and the document is precisely positioned at the corresponding document location. At the same time, the view state at that time is restored, including page number (such as page 5), scroll position (such as paragraph 8 of a Word document), zoom ratio, view mode (such as PPT presentation mode), etc., completely replicating the document viewing scenario at that point in the meeting. Second, presentation of document content change events: If the operation event associated with the interactive controls is of the type involving content changes, such as editing or annotation, the specific content changes caused by the operation event can be displayed. For example, a document snapshot (lightweight or full snapshot) before the operation and the document state after the operation can be retrieved and presented in a comparison format (such as split screen, highlighting the changed area); the details of the changes are clearly displayed, including text additions and deletions, format modifications, annotations, highlighting / underlining, etc. In addition, it can simultaneously display key information such as the operator and operation time, allowing users to intuitively understand the specific content and background of document changes.

[0073] Through the above processing, the abstract meeting minutes content is deeply bound to the concrete media stream and document status via a timeline, achieving the effect of "recreating a complete meeting scene by clicking a control." Compared to the inefficient method of manually searching videos and flipping through documents in traditional technologies, this step 204, through automated linkage and backtracking, allows users to quickly verify the basis of discussions and trace the source of document changes without additional operations, greatly improving the reuse efficiency and traceability accuracy of meeting information, thereby significantly enhancing the user experience.

[0074] Furthermore, in one embodiment, in response to detecting an operation event of a document used in the meeting, a document snapshot corresponding to the operation event is obtained. The document snapshot at least includes document state information of the document at the time the operation event occurred. In this embodiment, controlling the document to present a corresponding document state based on the operation event associated with the interactive control includes: obtaining a corresponding document snapshot based on the operation event associated with the interactive control; and controlling the document to present a corresponding document state based on the obtained document snapshot.

[0075] This embodiment provides a method for accurately and efficiently restoring the document state at the time of an operation event by taking a document snapshot when the operation event occurs. This avoids deviations in state restoration caused by subsequent document modifications or real-time loading, and further ensures the accuracy and stability of the linkage backtracking.

[0076] Specifically, when any operation event (such as page switching, editing, annotation, etc.) of a document is detected during the meeting, a document snapshot is generated synchronously. In addition, extra timed snapshot generation logic can be set to automatically generate document snapshots at preset time intervals (such as 30 seconds or 1 minute). This can compensate for potential omissions in snapshots triggered by key operation nodes, ensuring continuous and complete recording of document status during the meeting. This complements the snapshots triggered by key operation nodes, jointly constructing a comprehensive and time-series-complete document status traceability chain.

[0077] For example, the generated document snapshot includes at least the document state information at the time the operation event occurred. This may include a snapshot ID, a timestamp of the corresponding operation event, the current page number, scroll position, zoom level, view mode, and other state descriptions, as well as content carriers such as page screenshots and visible area text / charts (a lightweight snapshot or a full snapshot can be selected depending on the scenario). The execution entity of this application embodiment binds and stores each document snapshot with its corresponding operation event (here, "corresponding" can refer to a precise point-in-time correspondence or a time range-level match), ensuring that the corresponding document snapshot can be quickly indexed through the operation event, forming a one-to-one or one-to-many mapping of "operation event - document snapshot" (the same operation event can correspond to multiple document snapshots).

[0078] After the user triggers the interactive control, the execution entity in this embodiment reads the associated operation event information (such as operation event ID and timestamp) carried by the interactive control, and uses this as a search keyword to quickly locate the document snapshot corresponding to the operation event from the stored snapshot library; then, it parses the obtained document snapshot and extracts the complete document state information recorded therein, including but not limited to the target document's page number, scroll position, zoom level, view mode (such as PPT presentation mode), and visible content area. Based on the extracted snapshot state information, the following operations are performed to restore the document scene at the time of the operation event: Document location: Open the target document associated with the interactive control, and accurately locate it according to the page number and scroll position recorded in the snapshot to ensure that the displayed document area is consistent with that at the time of the meeting; State restoration: Synchronously adjust the document's zoom level and view mode to replicate the document viewing state in the meeting; Content display: If the snapshot contains a page screenshot or a complete copy of the content, the snapshot content can be directly displayed, or the corresponding area of ​​the document can be rendered based on the snapshot information to avoid state deviations caused by subsequent modifications to the document.

[0079] For example, suppose the user triggers the interaction control [00:54:32|Q3 Sales Report.pptx|Page 5|Edit Operation|Zhang San]. The corresponding document snapshot contains information such as "Page 5, Zoom Ratio 100%, View Mode Presentation Mode, Page Screenshot (including the title 'Sales Data' before editing)". After retrieving the document snapshot corresponding to this editing operation, "Q3 Sales Report.pptx" is automatically opened, page 5 is navigated to, the document zoom ratio is simultaneously adjusted to 100%, and the presentation mode is switched to ensure that the view state is consistent with that at the time of the meeting. If the user needs to view the original document content when the operation occurred, the page screenshot in the snapshot can be displayed directly, or the document state restored based on the snapshot can be highlighted to completely replicate the document scene when Zhang San performed the editing operation.

[0080] Since document snapshots are instantaneous "freeze-frames" of the event, unaffected by subsequent editing or deletion of the document, this embodiment uses document snapshots to present the document state, ensuring the accuracy of state restoration and guaranteeing that the document state seen by the user during rewind is completely consistent with the state at the time of the meeting. Simultaneously, document snapshots are pre-generated and structured for storage, eliminating the need for real-time calculation or loading of the original document data. State information is directly extracted and rendered from the snapshot, significantly reducing response time and improving document state presentation efficiency. Furthermore, it does not rely on the real-time availability of the original document file; even if the target document is modified or deleted, the historical state can still be independently presented through the snapshot, enhancing the robustness of the technical solution.

[0081] Figure 2 The process shown is in Figure 1 Based on the illustrated process, by responding to simple user triggers of interactive controls within the meeting minutes, media stream rewind and document state restoration can be automatically executed. This integrates the fragmented meeting information—audio and video streams, meeting minutes, and document operations—under a unified timeframe, eliminating the need for manual switching and searching across multiple platforms and files. Users can then immersively revisit the complete meeting scenario at any point in the discussion with a single click: hearing the speeches and discussions, seeing the document's specific page, view state, and even editing traces at that time. This completely solves the pain points of low efficiency and cumbersome operation in traditional information rewind methods, establishing a clear and traceable contextual relationship between meeting decisions and document changes. It significantly improves the quality of meeting knowledge retention and the efficiency of reusing team collaboration information, fundamentally changing the interactive paradigm of information recording and rewind in online meetings.

[0082] Figure 3 A flowchart illustrating an embodiment of another meeting minutes generation method provided in this application. Figure 3 The process shown is in Figure 1 Based on the illustrated process, the following steps are included: Step 301: During the meeting process, generate a first series of timestamps for the media stream of the meeting, and listen for operation events of documents used in the meeting, generate a second series of timestamps for the operation events, and generate the document editing history of the document based on the listened operation events.

[0083] Step 301 adds the action of "generating the document editing history of the document based on the monitored operation events" to build a data foundation for two-way traceability of meeting information, based on the above embodiment.

[0084] The document editing history is a specialized record formed by integrating and structuring all document operation events monitored in step 301. Its exemplary generation logic includes: reusing the operation events captured by the document operation monitoring module at the data level to form the document editing history; each record is bound to a second series of time stamps and includes the operation event type (insert / delete / modify / annotate, etc.), the specific content of the operation, the executor, the operation occurrence time (relative time stamp based on meeting T0), the associated second series of time stamp information, and the corresponding document snapshot ID. In terms of presentation, all document editing history records are arranged in chronological order of operation occurrence, forming a time-series history list that allows users to filter and view by time, operation type, editor, and other dimensions, ensuring that the operation trajectory is clearly traceable.

[0085] Step 302: Generate meeting minutes based on the media stream with the first series of time stamps.

[0086] Step 303: In the meeting minutes, configure an interactive control for at least one content unit. The interactive control is associated with at least one operation event and carries at least the time stamp information of the associated operation event.

[0087] Step 304: In response to the trigger operation of the interactive control, execute the first linkage backtracking operation.

[0088] For detailed explanations of steps 301 to 304, please refer to the above. Figure 1 and Figure 2 The relevant descriptions in the process shown will not be repeated here.

[0089] Step 305: In response to viewing the target operation event in the document editing history, execute the second linked backtracking operation.

[0090] The second linkage backtracking operation includes at least one of the following: locating and presenting content units associated with the target operation event in the meeting minutes based on the time stamp information of the target operation event; and causing the media stream to start displaying at the corresponding time point based on the time stamp information of the target operation event.

[0091] As can be seen from the above description, based on the document editing history formed in step 301, "reverse" linkage backtracking (i.e., second linkage backtracking) can be achieved. Here, "reverse" linkage backtracking means: taking the operation event in the document editing history as the starting point, and using the same-source alignment relationship of the time stamp, reversing and tracing back to the corresponding meeting scene information.

[0092] Specifically, the second linkage backtracking operation includes at least one of the following implementations: On the one hand, there is the positioning and presentation of meeting minutes content. In this embodiment, the executing entity reads the second series of timestamps of the target operation event and queries the meeting minutes based on these timestamps. Since each content unit (text paragraph, image, list item) of the minutes inherits the first series of timestamps that are from the same source as the second series of timestamps, the executing entity in this embodiment can accurately filter out content units whose time intervals contain these timestamps (e.g., when the target operation event timestamp is [00:54:32], it matches the paragraphs in the minutes from [00:54:20-00:54:45]), and then automatically jumps to the content unit, highlighting the paragraph and its surrounding context (e.g., speaker information, discussion topic), allowing users to intuitively see the meeting discussion content corresponding to the operation.

[0093] On the other hand, there is the precise positioning and display of the media stream. The execution entity of this application, based on the second series of time markers of the target operation event, locates the corresponding time point of the meeting media stream and starts the display. If the media stream contains audio, it will automatically jump to that moment to play the meeting recording, restoring the discussion voices of the participants around the operation at that time (such as the dialogue between Zhang San explaining the reasons for the modification and Li Si supplementing the suggestions). If it contains screen sharing video, the video screen at that moment will be played synchronously, replicating the screen scene when the operation was executed, forming a multi-dimensional linkage scene restoration with audio, video and minutes content.

[0094] For example, when a user clicks on the target operation event "00:54:32 | Zhang San | Editing operation: Change the title 'Sales Data' to 'Q3 Regional Sales Data Comparison'" in the document editing history, the execution entity in this embodiment can simultaneously execute two actions: First, automatically open the corresponding meeting minutes and highlight the content unit "Li Si: The title here needs to be changed to better fit the theme of regional data comparison" within the time interval [00:54:20-00:54:45]; Second, the meeting audio jumps to 00:54:32, plays Zhang San's discussion voice "I changed the title to 'Q3 Regional Sales Data Comparison' so that the differences between regions can be more clearly reflected", and simultaneously plays the screen sharing video at that time to display the real-time operation screen of the title modification.

[0095] Figure 3 The process shown is in Figure 1 and Figure 2 Based on the illustrated process, a reverse correlation and backtracking mechanism is constructed by generating and utilizing document editing history, linking document operations to meeting scenarios. This mechanism uses shared timestamps as a bridge to closely link discrete document editing events with meeting minutes content and media streams in the time dimension, thereby achieving bidirectional traceability of meeting information.

[0096] Based on any of the above embodiments, this application also provides the following embodiments: In one embodiment, configuring interactive controls for at least one content unit in the meeting minutes content includes: in response to determining that the same content unit in the meeting minutes content is associated with operation events for multiple different documents, configuring multiple interactive controls for the content unit, each interactive control being associated with a different document and a corresponding operation event.

[0097] This embodiment addresses the issue of association and interaction when the same minutes content unit is associated with operation events of multiple different documents in a meeting, and further improves the linkage and backtracking capabilities in multi-document scenarios, thus meeting the needs of cross-discussion of multiple materials in actual meetings.

[0098] Specifically, in typical scenarios such as enterprise solution reviews and cross-departmental collaborative meetings, participants often need to open multiple documents simultaneously (such as sales data reports, marketing plans, product design drafts, etc.) and discuss the content of these documents in a cross-cutting manner. In this case, the same content unit in the meeting minutes (such as a discussion conclusion or a to-do item) may correspond to the operation events of multiple different documents within the same time interval. For example, when discussing "adjusting regional distribution strategies," participants may scroll through the regional data page of "Q3 Sales Report.pptx" and add annotations to "2024 Marketing Plan.docx." These two operation events, belonging to different documents, are both directly related to the discussion content.

[0099] To address this, the technical solution provided in the above embodiments is applied to embed two interactive controls into the content unit "Wang Wu: Adjusting the East China Region's Placement Strategy Based on Sales Data and Promotion Plan" in the meeting minutes. The information associated with and carried by these two interactive controls is as follows: [10:52|Q3 Sales Report.pptx|Page 8|Scroll Operation] and [10:58|2024 Marketing Promotion Plan.docx|Page 3|Annotation Operation]. The former indicates that at 10:52, the participant scrolls "Q3 Sales Report.pptx" to page 8 (to view East China region sales data); the latter indicates that at 10:58, the participant adds the annotation "Increase placement in East China region by 20%" on page 3 of "2024 Marketing Promotion Plan.docx".

[0100] In the above embodiments, when there are operation events for at least two different documents, each interactive control is bound to the operation event of one document and carries corresponding information independently. For example, multiple interactive controls are presented in the minutes in order of document discussion or correlation strength, and can be clearly distinguished by document names, exclusive icons, etc., to avoid confusion. When a user triggers any interactive control, targeted actions will be executed according to the information carried by the control. For example, clicking the control corresponding to "Q3 Sales Report.pptx" will automatically open the document and position it on page 8, restoring the previous scroll position and view state, while simultaneously playing the meeting audio from 10:50 to 11:05; clicking the control corresponding to "2024 Marketing Plan.docx" will jump to the annotation position on page 3 of the document, synchronously playing the audio for the corresponding time period.

[0101] This embodiment overcomes the limitations of single-document association, adapts to actual meeting scenarios involving parallel discussions of multiple documents, ensures that the operation events of each document can be accurately bound to the minutes content, and effectively avoids information loss during multi-document discussions.

[0102] In one embodiment, during the meeting, based on the audio in the media stream, a speaker is identified using voiceprint recognition technology, and the start and end timestamps of the speaker's corresponding speaking segment are determined. Based on the start and end timestamps, the speaker's identifier is associated with the corresponding content unit in the meeting minutes. Further, in this embodiment, in response to a trigger operation on an interactive control, a first linked backtracking operation is executed, and the speaker's identifier associated with the content unit corresponding to the interactive control is displayed on the meeting minutes display interface.

[0103] This embodiment adds a speaker dimension to the three-dimensional linkage of "meeting minutes - documents used in the meeting - meeting media stream". The core logic is to bind speaker information with minutes content and document operations through voiceprint recognition technology, so that the linkage and retrospection can not only restore "discussion content + document status", but also clarify "who participated in the discussion", further improving the complete restoration of the meeting scene and enhancing the retrospection experience of multi-person meetings.

[0104] Specifically, during the meeting, based on the recorded media stream audio, the audio data is analyzed in real time using voiceprint recognition technology. On the one hand, the speaker's identity is accurately identified (generating a speaker identifier, such as the name "Zhang San", user ID, etc.). On the other hand, the start and end time markers of each speaker's corresponding speech segment are determined (from the same source as the first series of time markers, based on the relative time of the meeting start time T0, such as [10:35:00-10:38:20]), ensuring that the time range of the speech segment is consistent with the timeline of the media stream and document operation events.

[0105] Since each content unit of the meeting minutes (such as a single discussion text) originates from a specific time interval of the media stream, the start and end times of the speaker's speech segment can be matched with the time window corresponding to the content unit. If the two time ranges overlap or intersect (such as the speech segment [10:35:00-10:38:20] overlapping with the content unit's time window [10:35-10:39]), then the speaker's identifier is associated with that content unit. This association reflects the "composition of the participants in the discussion of this part (content unit)," rather than just the speaker at a single moment.

[0106] After the above association is completed, the speaker's identifier will serve as a key piece of information. When the user triggers the interactive control to perform the first linked retrospective operation, the associated speaker's identifier can also be dynamically displayed on the meeting minutes interface (e.g., next to the content unit, above the player, or in a separate information bar). (For example, "Participants: Zhang San, Li Si" will be listed as a label at the top of the content unit.)

[0107] This approach allows users not only to hear recordings of historical discussions and view the original documents, but also to clearly see which participants contributed to each segment of the minutes (each discussion point). It enables them to quickly grasp the scope of participation and responsible parties for different topics, recreating the full picture of the collaborative process during the meeting. This further enhances the traceability of meeting decisions and the visibility of collaborative contributions, representing a significant step forward in improving the structured depth of meeting knowledge, the traceability of collaboration, and the effectiveness of debriefing.

[0108] Figure 4 This is a block diagram illustrating an embodiment of a meeting minutes generation device provided in this application. Figure 4 As shown, the device includes: The first time stamp module 41 is used to generate a first series of time stamps for the media stream of the conference during the conference process; Document operation monitoring module 42 is used to monitor operation events of documents used in the meeting and generate a second series of timestamps for the operation events; The meeting minutes generation module 43 is used to generate meeting minutes content based on the media stream with the first series of time stamps; The document operation association module 44 is used to configure an interactive control for at least one content unit in the meeting minutes content. The interactive control is associated with at least one operation event and carries at least the time stamp information of the associated operation event.

[0109] In one possible implementation, the device further includes: The first linkage backtracking module is used to execute the first linkage backtracking operation in response to the trigger operation of the interactive control; The first linkage backtracking operation includes: causing the media stream to start displaying at the corresponding time point according to the time stamp information carried by the interactive control, and controlling the document to present the corresponding document state according to the operation events associated with the interactive control.

[0110] In one possible implementation, the device further includes: A document snapshot generation module is used to obtain a document snapshot corresponding to the operation event in response to listening to an operation event of a document used in the meeting. The document snapshot at least includes the document state information of the document at the time the operation event occurs. The first linkage backtracking module controls the document to present a corresponding document state based on the operation events associated with the interactive control, including: Based on the operation events associated with the interactive control, obtain the corresponding document snapshot; Based on the obtained document snapshot, control the document to present the corresponding document state.

[0111] In one possible implementation, the first linkage backtracking module controls the document to present a corresponding document state, including: Locate and display the document at the corresponding document position; And / or, presenting document content change events triggered by operation events corresponding to the document snapshot.

[0112] In one possible implementation, the device further includes: The editing history generation module is used to generate the document editing history of the document based on the monitored operation events; The second linkage backtracking module is used to execute a second linkage backtracking operation in response to a viewing operation of the target operation event in the document editing history. The second linkage backtracking operation includes at least one of the following: Based on the time stamp information of the target operation event, locate and present the content unit associated with the target operation event in the meeting minutes content; Based on the time stamp information of the target operation event, the media stream is displayed at the corresponding time point.

[0113] In one possible implementation, the device further includes: The highlighting module is used to highlight the content unit corresponding to the current display period in the display interface of the meeting minutes while displaying the media stream.

[0114] In one possible implementation, the document operation association module 44 is used to determine associated operation events for content units in the meeting minutes content in the following manner: Based on the media stream with the first series of time stamps, determine the time window corresponding to the content unit in the meeting minutes content; At least one operation event whose time stamp falls within the time window is identified as an operation event associated with the content unit.

[0115] In one possible implementation, the document operation association module 44 determines at least one operation event whose timestamp falls within the time window as an operation event associated with the content unit, including: Operation events whose timestamps fall within the time window are identified as candidate operation events. Determine the semantic relevance between the candidate operation events and the content unit, and based on the semantic relevance, determine the operation events associated with the content unit from the candidate operation events.

[0116] In one possible implementation, the device further includes: The speaker association module is used to identify speakers based on the audio in the media stream using voiceprint recognition technology during the meeting process, and to determine the start and end time markers of the speaking segments corresponding to the speaker. Based on the start and end time markers, the speaker identifier is associated with the corresponding content unit in the meeting minutes.

[0117] In one possible implementation, the device further includes: The speaker display module is used to respond to the trigger operation of the interactive control, perform a first linkage backtracking operation, and display the speaker identifier associated with the content unit corresponding to the interactive control in the display interface of the meeting minutes content.

[0118] In one possible implementation, the document operation association module 44 configures interactive controls for at least one content unit in the meeting minutes content, including: In response to determining that the same content unit in the meeting minutes is associated with operation events for multiple different documents, multiple interactive controls are configured for the content unit, each of which is associated with a different document and a corresponding operation event.

[0119] like Figure 5 As shown in the figure, this application provides an electronic device, including a processor 111, a communication interface 112, a memory 113, and a communication bus 114, wherein the processor 111, the communication interface 112, and the memory 113 communicate with each other through the communication bus 114. Memory 113 is used to store computer programs; In one embodiment of this application, when the processor 111 executes the program stored in the memory 113, it implements the meeting minutes generation method provided in any of the foregoing method embodiments, including: During the meeting, a first series of timestamps are generated for the media stream of the meeting; operation events using documents in the meeting are listened for, and a second series of timestamps are generated for the operation events; Based on the media stream with the first series of time stamps, generate meeting minutes content; In the meeting minutes, at least one content unit is configured with an interactive control, which is associated with at least one operation event and carries at least the time stamp information of the associated operation event.

[0120] This application also provides a computer-readable storage medium storing a computer program thereon, which, when executed by a processor, implements the steps of the meeting minutes generation method provided in any of the foregoing method embodiments.

[0121] The device embodiments described above are merely illustrative. The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the modules can be selected to achieve the purpose of this embodiment according to actual needs.

[0122] Through the above description of the embodiments, those skilled in the art can clearly understand that each embodiment can be implemented using software plus a general-purpose hardware platform, or of course, using hardware. Based on this understanding, the above technical solutions, in essence or the parts that contribute to the related technology, can be embodied in the form of a software product. This computer software product can be stored in a computer-readable storage medium, such as ROM / RAM, magnetic disk, optical disk, etc., and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute the methods described in the various embodiments or some parts of the embodiments.

[0123] It should be understood that the terminology used herein is for the purpose of describing particular exemplary embodiments only and is not intended to be limiting. Unless the context clearly indicates otherwise, the singular forms “a,” “an,” and “described” as used herein may also include the plural forms. The terms “comprising,” “including,” “containing,” and “having” are inclusive and therefore indicate the presence of the stated features, steps, operations, elements, and / or components, but do not exclude the presence or addition of one or more other features, steps, operations, elements, components, and / or combinations thereof. The method steps, processes, and operations described herein are not construed as requiring them to be performed in a particular order described or illustrated unless the order of performance is explicitly indicated. It should also be understood that additional or alternative steps may be used.

[0124] The above description is merely a specific embodiment of this application, enabling those skilled in the art to understand or implement this application. Various modifications to these embodiments will be readily apparent to those skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of this application. Therefore, this application is not to be limited to the embodiments shown herein, but is to be accorded the widest scope consistent with the principles and novel features claimed herein.

Claims

1. A method for generating meeting minutes, characterized in that, The method includes: During the meeting, a first series of timestamps are generated for the media stream of the meeting; operation events using documents in the meeting are listened for, and a second series of timestamps are generated for the operation events; Based on the media stream with the first series of time stamps, generate meeting minutes content; In the meeting minutes, at least one content unit is configured with an interactive control, which is associated with at least one operation event and carries at least the time stamp information of the associated operation event.

2. The method according to claim 1, characterized in that, The method further includes: In response to a trigger operation on the interactive control, a first linkage backtracking operation is executed; The first linkage backtracking operation includes: causing the media stream to start displaying at the corresponding time point according to the time stamp information carried by the interactive control, and controlling the document to present the corresponding document state according to the operation events associated with the interactive control.

3. The method according to claim 2, characterized in that, The method further includes: In response to detecting an operation event of a document used in the meeting, a document snapshot corresponding to the operation event is obtained, the document snapshot containing at least the document state information of the document at the time the operation event occurred; The step of controlling the document to present a corresponding document state based on the operation events associated with the interactive control includes: Based on the operation events associated with the interactive control, obtain the corresponding document snapshot; Based on the obtained document snapshot, control the document to present the corresponding document state.

4. The method according to claim 3, characterized in that, The control of the document to present the corresponding document state includes: Locate and display the document at the corresponding document position; And / or, presenting document content change events triggered by operation events corresponding to the document snapshot.

5. The method according to claim 1, characterized in that, The method further includes: Based on the monitored operation events, generate the document's document editing history; In response to viewing a target operation event in the document editing history, a second linked backtracking operation is executed, the second linked backtracking operation including at least one of the following: Based on the time stamp information of the target operation event, locate and present the content unit associated with the target operation event in the meeting minutes content; Based on the time stamp information of the target operation event, the media stream is displayed at the corresponding time point.

6. The method according to claim 2 or 5, characterized in that, The method further includes: While displaying the media stream, the content unit corresponding to the current display period is highlighted in the display interface of the meeting minutes content.

7. The method according to claim 1, characterized in that, The associated operation events for the content units in the meeting minutes are determined in the following manner: Based on the media stream with the first series of time stamps, determine the time window corresponding to the content unit in the meeting minutes content; At least one operation event whose time stamp falls within the time window is identified as an operation event associated with the content unit.

8. The method according to claim 7, characterized in that, The operation event that makes the timestamp fall within the time window is determined to be an operation event associated with the content unit, including: Operation events whose timestamps fall within the time window are identified as candidate operation events. Determine the semantic relevance between the candidate operation events and the content unit, and based on the semantic relevance, determine the operation events associated with the content unit from the candidate operation events.

9. The method according to claim 1, characterized in that, The method further includes: During the meeting, based on the audio in the media stream, the speaker is identified using voiceprint recognition technology, and the start and end time markers of the speaker's corresponding speaking segment are determined. Based on the start and end time markers, the speaker identifier is associated with the corresponding content unit in the meeting minutes.

10. The method according to claim 9, characterized in that, The method also includes In response to the triggering operation of the interactive control, a first linkage backtracking operation is performed, and the speaker's identifier associated with the content unit corresponding to the interactive control is displayed in the display interface of the meeting minutes content.

11. The method according to claim 1, characterized in that, The provision of interactive controls for at least one content unit in the meeting minutes includes: In response to determining that the same content unit in the meeting minutes is associated with operation events for multiple different documents, multiple interactive controls are configured for the content unit, each of which is associated with a different document and a corresponding operation event.

12. A meeting minutes generation device, characterized in that, The device includes: The first time stamp module is used to generate a first series of time stamps for the media stream of the conference during the conference process; A document operation monitoring module is used to monitor operation events of documents used in the meeting and generate a second series of timestamps for the operation events. The meeting minutes generation module is used to generate meeting minutes content based on the media stream with the first series of time stamps; The document operation association module is used to configure interactive controls for at least one content unit in the meeting minutes content. The interactive controls are associated with at least one operation event and at least carry the time stamp information of the associated operation event.

13. An electronic device, characterized in that, include: A processor and a memory, the processor being configured to execute a meeting minutes generation program stored in the memory to implement the meeting minutes generation method according to any one of claims 1-11.

Images