Video editing method and device

Abstract

The embodiment of the invention provides a video editing method. The video editing method comprises the following steps: acquiring text description content of a target video; determining a plurality of candidate segments from the target video according to the text description content; a plurality of intelligent agents are adopted to score the candidate segments, a scoring result is obtained, and different intelligent agents carry out different scores on the candidate segments; determining a plurality of target fragments from the plurality of candidate fragments according to a scoring result; and editing the target clip, and outputting the edited video clip. According to the technical scheme of the embodiment of the invention, the operation of converting live broadcast to video-on-demand can be automatically completed, and the video editing efficiency is improved, so that a large amount of live broadcast content can be dealt with, and the demand of audiences for watching video-on-demand as soon as possible is met.

Description

Technical Field

[0001] The embodiments of the present application relate to the field of artificial intelligence technology, and in particular, to a video editing method, apparatus, computer equipment, computer-readable storage medium, and computer program product. Background Art

[0002] With the development of information technology and the increasing demand for content from viewers, how to more efficiently use video resources and quickly launch valuable content has become an important issue. One of them is to convert live broadcast to on-demand. After the live broadcast ends, viewers usually want to watch on-demand content as soon as possible so that they can find the clips they are interested in and review them. In order to enable viewers to quickly find the parts they are interested in, it is usually necessary to split the highlights of the live broadcast into multiple clips and release them so that viewers can quickly select a video clip to play.

[0003] The current live broadcast to VOD operation usually relies on manual work. When processing the live broadcast to VOD conversion, manual work needs to carefully watch and screen the video content, and then perform operations such as editing and splicing. This makes manual editing not only time-consuming and labor-intensive, but also inefficient. It is difficult to cope with a large amount of live content. The whole process is long and it is difficult to meet the audience's needs to watch VOD as soon as possible.

[0004] It should be noted that the above content is not necessarily prior art, nor is it intended to limit the scope of patent protection of this application. Summary of the invention

[0005] The embodiments of the present application provide a video editing method, apparatus, computer device, computer-readable storage medium, and computer program product to solve or alleviate one or more of the technical problems raised above.

[0006] One aspect of an embodiment of the present application provides a video editing method, the method comprising: Get the text description content of the target video; Determining a plurality of candidate segments from the target video according to the text description content; Using a plurality of intelligent agents to score the plurality of candidate segments to obtain scoring results, wherein different intelligent agents give different scores to the candidate segments; Determining a plurality of target segments from the plurality of candidate segments according to the scoring result; The target segment is clipped and the clipped video segment is output.

[0007] Optionally, determining a plurality of candidate segments from the target video according to the text description content includes: A large language model agent is used to select multiple segments from the target video according to the text description content, and generate structured information for each of the segments to obtain multiple candidate segments containing the structured information, wherein the structured information includes at least part of the segment number, starting position, ending position, segment title and segment content.

[0008] Optionally, the adopting several agents to score the multiple candidate segments includes: Using several intelligent agents to score the structured information of the candidate fragment; and / or obtaining the text fragment corresponding to the candidate fragment from the text description content, and using the several intelligent agents to score the text fragment.

[0009] Optionally, determining a plurality of candidate segments from the target video according to the text description content includes: Using a topic segmentation agent to determine different topics in the text description content, and dividing the target video into multiple topic segments according to the different topics; A content selection agent is used to determine multiple candidate segments from the multiple topic segments according to the text description content.

[0010] Optionally, the adopting several agents to score the multiple candidate segments includes: Using a content quality assessment agent to score the multiple candidate segments to obtain a content quality score for each candidate segment; Using an attractiveness evaluation agent to score the multiple candidate segments to obtain an attractiveness score for each of the candidate segments; Using a resonance evaluation agent to score the multiple candidate segments to obtain a resonance score for each candidate segment; The multiple candidate segments are scored using a content compliance assessment agent to obtain a content compliance score for each candidate segment.

[0011] Optionally, determining a plurality of target segments from the plurality of candidate segments according to the scoring result includes: Using a comprehensive scoring agent to determine a comprehensive score for each candidate segment based on the content quality score, the attractiveness score, the resonance score, and the content compliance score; The comprehensive scoring agent is used to determine the top N candidate segments as the target segments according to the comprehensive scores, where N is a positive integer.

[0012] Optionally, obtaining the text description content of the target video includes: Get the voice content of the target video; Automatic speech recognition is used to convert the speech content into text description content of the target video.

[0013] Another aspect of an embodiment of the present application provides a video editing device, the device comprising: An acquisition module is used to obtain the text description content of the target video; A first determination module, configured to determine a plurality of candidate segments from the target video according to the text description content; A scoring module, configured to use a plurality of intelligent agents to score the plurality of candidate segments to obtain scoring results, wherein different intelligent agents give different scores to the candidate segments; A second determination module, configured to determine a plurality of target segments from the plurality of candidate segments according to the scoring result; The editing module is used to edit the target segment and output the edited video segment.

[0014] Another aspect of an embodiment of the present application provides a computer device, comprising: at least one processor; and a memory communicatively connected to the at least one processor; wherein: the memory stores instructions executable by the at least one processor, and the instructions are executed by the at least one processor to enable the at least one processor to execute the method described above.

[0015] Another aspect of an embodiment of the present application provides a computer-readable storage medium, wherein the computer-readable storage medium stores computer instructions, and when the computer instructions are executed by a processor, the method described above is implemented.

[0016] Another aspect of an embodiment of the present application provides a computer program product, including a computer program, which implements the method described above when executed by a processor.

[0017] The above technical solution adopted in the embodiment of the present application may have the following advantages: By obtaining the text description content of the target video, multiple candidate segments are determined from the target video according to the text description content, and several intelligent agents are used to score the multiple candidate segments to obtain scoring results. Different intelligent agents give different scores to the candidate segments, and several target segments are determined from the multiple candidate segments according to the scoring results. The target segments are edited and the edited video segments are output. Automatic video editing can be achieved and the efficiency of video editing can be improved, so that a large amount of live content can be handled and the audience's needs for watching on demand as soon as possible can be met. BRIEF DESCRIPTION OF THE DRAWINGS

[0018] The accompanying drawings exemplarily illustrate the embodiments and constitute a part of the specification, and together with the text description of the specification, are used to explain the exemplary implementation of the embodiments. The embodiments shown are for illustrative purposes only and do not limit the scope of the claims. In all drawings, the same reference numerals refer to similar but not necessarily identical elements.

[0019] Figure 1 The following schematically shows an operating environment diagram of a video editing method according to Embodiment 1 of the present application; Figure 2 The flowchart of the video editing method according to the first embodiment of the present application is schematically shown; Figure 3 Schematically shows Figure 2 Flow chart of sub-steps of step S200; Figure 4 Schematically shows Figure 2 Flow chart of sub-steps of step S202; Figure 5 Schematically shows Figure 2 Flow chart of sub-steps of step S204; Figure 6 Schematically shows Figure 2 Flow chart of sub-steps of step S206; Figure 7 The principle diagram of the video editing method according to the first embodiment of the present application is schematically shown; Figure 8 A block diagram schematically shows a video editing device according to the second embodiment of the present application; and Fig. 9 The hardware architecture diagram of the computer device according to the third embodiment of the present application is schematically shown. DETAILED DESCRIPTION

[0020] In order to make the purpose, technical solutions and advantages of the present application more clearly understood, the present application is further described in detail below in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present application and are not intended to limit the present application. Based on the embodiments in the present application, all other embodiments obtained by ordinary technicians in the field without making creative work are within the scope of protection of the present application.

[0021] It should be noted that the descriptions involving "first", "second", etc. in the embodiments of the present application are only for descriptive purposes and cannot be understood as indicating or implying their relative importance or implicitly indicating the number of technical features indicated. Therefore, the features defined as "first" and "second" may explicitly or implicitly include at least one of the features. In addition, the technical solutions between the various embodiments can be combined with each other, but they must be based on the ability of ordinary technicians in the field to implement them. When the combination of technical solutions is contradictory or cannot be implemented, it should be deemed that such combination of technical solutions does not exist and is not within the scope of protection required by this application.

[0022] In the description of the present application, it should be understood that the numerical labels before the steps do not indicate the order in which the steps are executed, but are only used to facilitate the description of the present application and to distinguish each step, and therefore should not be understood as a limitation on the present application.

[0023] First, the following terms are explained: Agent: It is an important concept in the field of artificial intelligence (AI). It refers to an intelligent entity that can autonomously perceive the environment, make decisions and perform actions. It can be a program, a system or a robot. The core of the agent is artificial intelligence algorithms, including machine learning, deep learning, reinforcement learning, neural networks and other technologies. Through these algorithms, the agent can learn from a large amount of data and improve its own performance, and continuously optimize its own decisions and behaviors. The agent can also make flexible adjustments according to environmental changes to adapt to different scenarios and tasks.

[0024] Large Language Model Agents (LLM Agents): An intelligent entity that can perceive the environment, make decisions, and execute actions. With a large language model as its core brain, it has the ability to gradually complete a given goal by thinking independently and calling tools. LLM agents not only respond to user query instructions to generate tasks, but also decompose complex tasks and make autonomous decisions in each sub-step, solving problems without user participation.

[0025] Live to VOD: It is the process of synchronously recording the live stream as VOD video and supports a series of subsequent operations. Specifically, it supports operations such as media resource management, media processing (such as transcoding, content review, smart cover and other AI processing), content production (such as cloud editing), etc. By converting live to VOD, the content in the live broadcast can be recorded and converted into VOD video so that the audience can watch it at any time after the live broadcast ends.

[0026] Automatic Speech Recognition (ASR) is a technology that converts human speech into text. It can process real-time audio streams or recorded audio files and convert them into readable text form.

[0027] Secondly, in order to facilitate those skilled in the art to understand the technical solutions provided in the embodiments of the present application, the relevant technologies are described below: In the related art, live broadcast to VOD conversion is usually done manually. Manual processing of live broadcast to VOD conversion is time-consuming, labor-intensive, inefficient, and difficult to cope with a large amount of live content. The whole process is long and prone to errors, making it difficult to meet the audience's needs for watching VOD as quickly as possible.

[0028] To this end, the embodiment of the present application provides a video editing technical solution. In this technical solution, the operation of converting live broadcast to on-demand broadcast can be completed automatically, and the efficiency of video editing can be improved, so that a large amount of live broadcast content can be handled and the audience's needs for watching on-demand broadcast as soon as possible can be met. See below for details.

[0029] Finally, for ease of understanding, an exemplary operating environment is provided below.

[0030] like Figure 1 As shown, the environment diagram includes a service platform 2, a network 4, and a client 6, wherein: The service platform 2 may be composed of a single or multiple computing devices. The multiple computing devices may include virtualized computing instances. Virtualized computing instances may include virtual machines, such as simulations of computer systems, operating systems, servers, etc. The computing device may load a virtual machine based on a virtual image and / or other data defining specific software (e.g., operating system, dedicated application, server) for simulation. As the demand for different types of processing services changes, different virtual machines may be loaded and / or terminated on one or more computing devices. A hypervisor may be implemented to manage the use of different virtual machines on the same computing device.

[0031] The service platform 2 may be configured to communicate with the client 6 or the like via a network 4. The network 4 includes various network devices, such as routers, switches, multiplexers, hubs, modems, bridges, repeaters, firewalls, proxy devices, and / or the like. The network 4 may include physical links, such as coaxial cable links, twisted pair cable links, optical fiber links, combinations thereof, and the like, or wireless links, such as cellular links, satellite links, Wi-Fi links, and the like.

[0032] The service platform 2 can provide storage, reading, writing, querying, deleting and other services, such as providing a service for reading clipped videos for the client.

[0033] The client 6 may be an electronic device running an operating system such as Windows, Android™ or iOS, such as a smart phone, a tablet device, a laptop computer, a virtual reality device, a gaming device, a set-top box, a vehicle terminal, or a smart TV. Based on the above operating system, various applications may be run, such as an application for viewing a clipped video.

[0034] The client 6 may provide / configure a user access page for manipulating the service platform 2 or uploading an object, etc.

[0035] It should be noted that the above devices are exemplary, and the number and type of devices are adjustable in different scenarios or according to different needs.

[0036] The following uses the service platform as the execution subject to introduce the technical solution of the present application through multiple embodiments. It should be noted that these embodiments can be implemented in a variety of different forms and should not be interpreted as being limited to the embodiments described here.

[0037] Embodiment 1 Figure 2 The flowchart of the video editing method according to the first embodiment of the present application is schematically shown.

[0038] like Figure 2 As shown, the video editing method may include steps S200 to S208, wherein: Step S200, obtaining text description content of the target video.

[0039] Step S202: determining a plurality of candidate segments from the target video according to the text description content.

[0040] Step S204, using a plurality of intelligent agents to score the plurality of candidate segments to obtain scoring results, wherein different intelligent agents give different scores to the candidate segments.

[0041] Step S206: determining a number of target segments from the plurality of candidate segments according to the scoring results.

[0042] Step S208, clipping the target segment and outputting the clipped video segment.

[0043] The video editing method provided in the present embodiment obtains the text description content of the target video, determines multiple candidate segments from the target video according to the text description content, uses several intelligent agents to score the multiple candidate segments, obtains scoring results, and different intelligent agents give different scores to the candidate segments. According to the scoring results, several target segments are determined from the multiple candidate segments, the target segments are edited, and the edited video segments are output. Automatic video editing can be achieved, and the efficiency of video editing is improved, so that a large amount of live content can be handled and the audience's needs for watching on demand as soon as possible can be met.

[0044] The following combination Figure 2 , each step in steps S200~S208 and other optional steps are explained in detail.

[0045] Step S200 , get the text description content of the target video.

[0046] The text description content of the target video may include text content such as subtitles, bullet comments, titles, and content descriptions, and may also be text description content obtained by recognizing the audio content of the target video.

[0047] In an optional embodiment, if Figure 3 As shown, step S200 may include steps S600 to S602: Step S600, obtaining the voice content of the target video.

[0048] Step S602: Automatic speech recognition is used to convert the speech content into text description content of the target video.

[0049] The voice content of the target video may include but is not limited to narrative content, expressive content and interactive content. Narrative content may include, for example, storytelling, commentary of documentaries or teaching videos, news reports, etc. Expressive content may include, for example, expression of opinions, poetry recitation and song singing, etc. Interactive content may include, for example, interviews, discussions, classroom teaching and live game broadcasts, etc.

[0050] After the text description content of the target video is obtained according to the voice content conversion, the voice content is converted into text content through automatic speech recognition, thereby obtaining the text description content of the target video. The text description content of the target video can be aligned with the timestamp of the video frame, so that the target video can be subsequently clipped according to the alignment relationship between the text description content and the timestamp of the video frame.

[0051] In this optional embodiment, by acquiring the voice content of the target video and using automatic speech recognition to convert the voice content into text description content of the target video, the text description content of the target video can be obtained based on the voice content of the target video. At the same time, by using automatic speech recognition technology to convert speech to text, the accuracy of the text description content can be improved.

[0052] Step S202 , multiple candidate segments are determined from the target video according to the text description content.

[0053] In an optional embodiment, step S202 may further include: using a large language model agent to select multiple segments from the target video according to the text description content, and generate structured information for each segment to obtain multiple candidate segments containing structured information, wherein the structured information includes the segment number, starting position, ending position, segment title and at least part of the segment content.

[0054] Specifically, a large language model agent can be used to perform semantic analysis, keyword recognition, sentiment analysis and other operations on the text description content, and multiple candidate segments can be determined from the target video according to the results of the operations. For example, by identifying whether the text description content contains keywords such as "key points", "key", "emphasis", "wonderful" and "high energy", the adjacent text description content of the keywords can be further analyzed (such as analyzing the context, keyword frequency changes or semantic jump points, etc.) to determine the content most relevant to the corresponding keywords, thereby determining the starting point and ending point corresponding to the segment; then, according to the corresponding relationship between the starting point and the ending point and the timestamp of the video frame, the corresponding video segment is determined as a candidate segment. For another example, by performing sentiment analysis on the text description content, if the emotion corresponding to the text description content is passionate, cheerful, or joyful, the text description content of the corresponding part can be further analyzed (such as analyzing the context, keyword frequency changes or semantic jump points, etc.) to determine the content most relevant to the corresponding emotional expression, thereby determining the starting point and ending point corresponding to the segment; then, according to the corresponding relationship between the starting point and the ending point and the timestamp of the video frame, the corresponding video segment is determined as a candidate segment.

[0055] After selecting multiple segments, the structured information of each segment is generated according to the information of each segment, and multiple candidate segments containing structured information are obtained. Among them, after selecting multiple segments, a complete sentence or a portion with a specified number of words at the starting point of each segment can be used as the starting position, and a complete sentence or a portion with a specified number of words at the ending point of each segment can be used as the ending position; a large language model intelligent agent can be used to analyze the text description content of the current segment, generate the title and segment content of the current segment, and generate a unique segment number for each segment to facilitate subsequent identification and calling.

[0056] In this optional embodiment, a large language model agent is used to select multiple segments from the target video according to the text description content, and structured information of each segment is generated to obtain multiple candidate segments containing structured information, wherein the structured information includes the segment number, starting position, ending position, segment title and at least part of the information in the segment content. The advantage of the large language model agent in text analysis can be used to accurately select candidate segments suitable for conversion to on-demand, thereby improving the accuracy of the candidate segments. At the same time, the generated structured information can facilitate the evaluation of the corresponding candidate segments, thereby improving the convenience of subsequent evaluation.

[0057] In an optional embodiment, if Figure 4 As shown, step S202 may further include steps S300-S302: Step S300, using a topic segmentation agent to determine different topics in the text description content, and dividing the target video into multiple topic segments according to the different topics.

[0058] Step S302: A content selection agent is used to determine a plurality of candidate segments from a plurality of topic segments according to the text description content.

[0059] Specifically, the topic segmentation agent can use natural language processing technology to analyze the text description content of the target video, and identify different topics in the text description content according to the analysis results by analyzing the semantic structure, keywords, syntax and other information of the text description content. On the basis of identifying different topics in the text description content, the context, keyword frequency changes, semantic jumps and other methods can be further analyzed to determine the topic transition points in the text description content; after determining the topic transition points, the topic switching agent can divide the target video into multiple topic segments according to the positions of these topic transition points; and then the content selection agent is used to determine multiple candidate segments from multiple topic segments according to the text description content. When the content selection agent is used to determine multiple candidate segments from multiple topic segments according to the text description content, the starting point and the ending point of the candidate segments can be determined by a method similar to the topic segmentation agent dividing the topic segments. In specific operations, the topic segmentation agent can use a coarse-grained method to divide, while the content selection agent can use a fine-grained method to obtain the starting point and the ending point.

[0060] In this optional embodiment, different topics in the text description content are determined through a topic switching agent, and the target video is divided into multiple topic segments according to the different topics. A large language model is used to determine multiple candidate segments from the multiple topic segments according to the text description content according to the agent. The candidate segments can be determined on the basis of the division into topic segments, thereby further improving the accuracy of obtaining the candidate segments. At the same time, different agents are used to respectively divide the topic segments and determine the candidate segments, so that different agents can be responsible for different content, which is beneficial to improving the accuracy of the agent operation results.

[0061] Step S204 , using several intelligent agents to score multiple candidate segments to obtain scoring results, where different intelligent agents give different scores to the candidate segments.

[0062] Specifically, the service platform can use several agents to score each candidate segment, and the score corresponding to each agent can be an evaluation of whether each candidate segment meets one dimension of the essence segment, so that the scoring results of each candidate segment in multiple dimensions can be obtained, so as to more comprehensively evaluate whether the candidate segment can be used as the essence segment of the target video. The specific dimensions can be set according to actual needs and actual audiences, and no specific restrictions are made here.

[0063] In an optional embodiment, based on the aforementioned generation of candidate segments containing structured information, in step S204, using several agents to score multiple candidate segments may further include: using several agents to score the structured information of the candidate segments; and / or obtaining text segments corresponding to the candidate segments from the text description content, and using several agents to score the text segments.

[0064] That is, when several agents are used to score each candidate segment, the structured information and / or the corresponding text segment corresponding to the candidate segment are scored. The agent can use a large language model agent to accurately score the text by taking advantage of the large language model agent's analysis of the text.

[0065] It should be understood that when using an agent to score candidate segments, scoring the structured information and / or text segments of the candidate segments can make the agent consume less computing resources, increase the speed of the agent evaluation, and thus further increase the speed of live broadcast to on-demand conversion.

[0066] In an optional embodiment, if Figure 5 As shown, step S204 may further include steps S400 to S406: Step S400: Use a content quality assessment agent to score multiple candidate segments to obtain a content quality score for each candidate segment.

[0067] Step S402: Use an attractiveness evaluation agent to score multiple candidate segments to obtain an attractiveness score for each candidate segment.

[0068] Step S404: Use a resonance evaluation agent to score the multiple candidate segments to obtain a resonance score for each candidate segment.

[0069] Step S406: Use a content compliance assessment agent to score multiple candidate segments to obtain a content compliance score for each candidate segment.

[0070] The content quality assessment agent, the attractiveness assessment agent, the resonance assessment agent and the content compliance assessment agent can score the content quality of the text fragments, structured information and / or video content corresponding to the candidate fragments.

[0071] When evaluating text types (text fragments or structured information), the content quality assessment agent can evaluate the grammatical accuracy, semantic coherence, information accuracy, and clarity of expression of the text; for video types (video content), it can evaluate the resolution, clarity, color, fluency, contrast, etc. of the video. When scoring content quality, a score of 0-10 can be given.

[0072] The attractiveness evaluation agent can first determine the audience of the video clip, and then evaluate the attractiveness of the video clip to the audience. Specifically, it can be evaluated from the aspects of the novelty of the theme of the text type or video type, the value of the content, the story, the value fit, the attractiveness of the title, the cover of the video clip, the number of interactions, etc., to determine the popularity of the candidate clip among the audience. When scoring the attractiveness, a score of 0-10 can be given.

[0073] The resonance evaluation agent can first determine the audience of the video clip, and then evaluate the emotional arousal, emotional intensity, emotional resonance, values, narrative techniques, style, etc. of the text type or video type based on the audience's interests, values, and cultural background to determine the degree of resonance of the candidate clip among the audience. When scoring resonance, a score of 0-10 can be given.

[0074] The content compliance assessment agent can identify sensitive content of text or video types according to platform rules and relevant laws and regulations, and evaluate the content compliance based on the identification results. When scoring content compliance, it can give a score of 0-10 points, or a score of 0 (false) or 1 (true).

[0075] It should be noted that when evaluating candidate clips using content quality assessment agents, attractiveness assessment agents, resonance assessment agents, and content compliance assessment agents, the evaluation can be performed in a certain order (serial) or simultaneously (in parallel). Preferably, the evaluation can be performed in parallel to improve the efficiency of the evaluation.

[0076] In this optional embodiment, by using a content quality assessment agent, an attractiveness assessment agent, a resonance assessment agent, and a content compliance assessment agent to score the candidate segments respectively, the content quality score, attractiveness score, resonance score, and content compliance score of each candidate segment can be obtained. The candidate segments can be evaluated in multiple dimensions through the collaboration of multiple agents, which can not only enable a single agent to focus on a specific evaluation and improve the professionalism and accuracy of the evaluation, but also obtain accurate evaluation results through the mutual integration, supplementation, and correction of the agent evaluations; and the evaluation time can be shortened and the efficiency of the evaluation can be improved through parallel methods. In addition, it is also possible to increase or decrease the number of agents to evaluate the candidate segments according to actual needs, thereby improving the adaptability to task requirements and environments.

[0077] Step S206 , several target segments are determined from multiple candidate segments according to the scoring results.

[0078] In an optional embodiment, if Figure 6 As shown, step S206 may further include steps S500-S502: Step S500, using a comprehensive scoring agent to determine the comprehensive score of each candidate segment based on the content quality score, attractiveness score, resonance score and content compliance score.

[0079] Step S502: A comprehensive scoring agent is used to determine the top N candidate segments as target segments according to the comprehensive scores, where N is a positive integer.

[0080] The comprehensive scoring agent can pre-process the data format of each score, such as processing the decimal point of the score, so as to determine the uniformity and validity of the data format.

[0081] Specifically, a comprehensive scoring agent can be used to set a weight for the content quality score, attractiveness score, resonance score and content compliance score respectively, and then the comprehensive score of each candidate segment is calculated by weighted average. The candidate segments are then sorted according to their comprehensive scores, and the top N candidate segments are determined as target segments.

[0082] For example, the weights can be: content quality score 30%, attractiveness score 25%, resonance score 20%, content compliance score 25%, then the comprehensive score can be calculated by the following formula: Comprehensive score = (content quality score*0.3) + (attractiveness score*0.25) + (resonance score*0.2) + (content compliance score*0.25).

[0083] In this optional embodiment, a comprehensive scoring agent is used to determine the comprehensive score of each candidate segment based on the content quality score, attractiveness score, resonance score and content compliance score, and the top N candidate segments are determined as target segments based on the comprehensive scores. The scores can be combined to obtain an overall evaluation of the candidate segments, thereby facilitating the evaluation of the essence segments.

[0084] Step S208 , clip the target segment and output the clipped video segment.

[0085] Specifically, in the case where the evaluation is performed on the text type of the segment (structured information or text segment), the starting video frame and the ending video frame corresponding to the target segment can be determined based on the alignment relationship between the structured information of the target segment or the text segment and the timestamp of the video frame, and then clipped, and finally the clipped video segment can be output.

[0086] Please refer to Figure 7 , which is a schematic diagram of the principle of the video editing method. The figure evaluates the structural information of the candidate segments, which can generally include the following: 1. Obtain the voice content of the target video and use automatic speech recognition (ASR) to convert the voice content into text description content; 2. The topic segmentation agent uses natural language processing technology to analyze the semantic structure, keywords, syntax and other information of the text description content to identify different topics in the text description content; then, the topic segmentation agent determines the topic transition points in the text description content by analyzing the context, keyword frequency changes, semantic jumps and other methods, and divides the text description content according to the topic transition points to obtain multiple topic segments; 3. The content selection agent uses a method similar to the topic segmentation agent but with finer granularity to select multiple candidate segments from the topic segments and generate structured information of the candidate segments, including the segment number, start position, end position, title and content summary of the candidate segments, and obtain multiple structures (structured information) corresponding to the multiple candidate segments; 4. Use content quality assessment agent, attractiveness assessment agent, resonance assessment agent and content compliance assessment agent to evaluate the structure respectively, and obtain content quality score, attractiveness score, resonance score and content compliance score. Among them, multiple agents can use information pool to share information, and the shared information can include the score of a certain structure. The subsequent comprehensive scoring agent can also obtain the scores of multiple evaluation agents on a certain structure from the information pool to perform comprehensive scoring; 5. The comprehensive scoring agent combines the scoring results of each evaluation agent to obtain a comprehensive score for each candidate segment, sorts them according to the comprehensive score, determines N candidate segments for editing, and finally outputs the essence of the N clips.

[0087] Embodiment 2 Figure 8 The block diagram of a video editing device 700 according to the second embodiment of the present application is schematically shown. The device 700 can be divided into one or more program modules, one or more program modules are stored in a storage medium, and are executed by one or more processors to complete the embodiment of the present application. The program module referred to in the embodiment of the present application refers to a series of computer program instruction segments that can complete specific functions. The following description will specifically introduce the functions of each program module in this embodiment. As shown in the figure, the device 700 may include: an acquisition module 710, a first determination module 720, a scoring module 730, a second determination module 740 and an editing module 750, wherein: An acquisition module 710 is used to acquire text description content of a target video; A first determination module 720, configured to determine a plurality of candidate segments from the target video according to the text description content; A scoring module 730 is used to use a plurality of agents to score the plurality of candidate segments to obtain scoring results, wherein different agents give different scores to the candidate segments; A second determination module 740, configured to determine a plurality of target segments from the plurality of candidate segments according to the scoring result; The editing module 750 is used to edit the target segment and output the edited video segment.

[0088] In an optional embodiment, the first determining module 720 is further configured to: A large language model agent is used to select multiple segments from the target video according to the text description content, and generate structured information for each of the segments to obtain multiple candidate segments containing the structured information, wherein the structured information includes at least part of the segment number, starting position, ending position, segment title and segment content.

[0089] In an optional embodiment, the scoring module 730 is further configured to: Using several intelligent agents to score the structured information of the candidate fragment; and / or obtaining the text fragment corresponding to the candidate fragment from the text description content, and using the several intelligent agents to score the text fragment.

[0090] In an optional embodiment, the first determining module 720 is further configured to: Using a topic segmentation agent to determine different topics in the text description content, and dividing the target video into multiple topic segments according to the different topics; A content selection agent is used to determine multiple candidate segments from the multiple topic segments according to the text description content.

[0091] In an optional embodiment, the scoring module 730 is further configured to: Using a content quality assessment agent to score the multiple candidate segments to obtain a content quality score for each candidate segment; Using an attractiveness evaluation agent to score the multiple candidate segments to obtain an attractiveness score for each of the candidate segments; Using a resonance evaluation agent to score the multiple candidate segments to obtain a resonance score for each candidate segment; The multiple candidate segments are scored using a content compliance assessment agent to obtain a content compliance score for each candidate segment.

[0092] In an optional embodiment, the second determining module 740 is further configured to: Using a comprehensive scoring agent to determine a comprehensive score for each candidate segment based on the content quality score, the attractiveness score, the resonance score, and the content compliance score; The comprehensive scoring agent is used to determine the top N candidate segments as the target segments according to the comprehensive scores, where N is a positive integer.

[0093] In an optional embodiment, the acquisition module 710 is further configured to: Get the voice content of the target video; Automatic speech recognition is used to convert the speech content into text description content of the target video.

[0094] Embodiment 3 Fig. 9The schematic diagram of the hardware architecture of a computer device 10000 suitable for implementing a video editing method according to the third embodiment of the present application is shown schematically. In some embodiments, the computer device 10000 may be a terminal device such as a smart phone, a wearable device, a tablet computer, a personal computer, a vehicle-mounted terminal, a game console, a virtual device, a workbench, a digital assistant, a set-top box, a robot, etc. In other embodiments, the computer device 10000 may be a rack server, a blade server, a tower server, or a cabinet server (including an independent server, or a server cluster composed of multiple servers), etc. Fig. 9 As shown, the computer device 10000 includes but is not limited to: a memory 10010, a processor 10020, and a network interface 10030 that can communicate with each other through a system bus. Among them: The memory 10010 includes at least one type of computer-readable storage medium, and the readable storage medium includes flash memory, hard disk, multimedia card, card-type memory (such as SD or DX memory), random access memory (RAM), static random access memory (SRAM), read-only memory (ROM), electrically erasable programmable read-only memory (EEPROM), programmable read-only memory (PROM), magnetic memory, magnetic disk, optical disk, etc. In some embodiments, the memory 10010 can be an internal storage module of the computer device 10000, such as a hard disk or memory of the computer device 10000. In other embodiments, the memory 10010 can also be an external storage device of the computer device 10000, such as a plug-in hard disk equipped on the computer device 10000, a smart memory card (Smart Media Card, SMC), a secure digital (Secure Digital, SD) card, a flash card (Flash Card), etc. Of course, the memory 10010 can also include both the internal storage module of the computer device 10000 and its external storage device. In this embodiment, the memory 10010 is generally used to store an operating system and various application software installed in the computer device 10000, such as program codes of a video editing method, etc. In addition, the memory 10010 can also be used to temporarily store various data that have been output or are to be output.

[0095] In some embodiments, the processor 10020 may be a central processing unit (CPU), a controller, a microcontroller, a microprocessor, or other chips. The processor 10020 is generally used to control the overall operation of the computer device 10000, such as performing control and processing related to data interaction or communication with the computer device 10000. In this embodiment, the processor 10020 is used to run the program code stored in the memory 10010 or process data.

[0096] The network interface 10030 may include a wireless network interface or a wired network interface, and the network interface 10030 is generally used to establish a communication link between the computer device 10000 and other computer devices. For example, the network interface 10030 is used to connect the computer device 10000 to an external terminal through a network, and to establish a data transmission channel and a communication link between the computer device 10000 and the external terminal. The network may be a wireless or wired network such as an intranet, the Internet, the Global System of Mobile communication (GSM), Wideband Code Division Multiple Access (WCDMA), 4G network, 5G network, Bluetooth, Wi-Fi, etc.

[0097] It should be pointed out that Fig. 9 Only a computer device having components 10010 - 10030 is shown, but it should be understood that implementation of all of the components shown is not a requirement, and more or fewer components may alternatively be implemented.

[0098] In this embodiment, the video editing method stored in the memory 10010 can also be divided into one or more program modules and executed by one or more processors (such as processor 10020) to complete the embodiment of the present application.

[0099] Embodiment 4 An embodiment of the present application further provides a computer-readable storage medium, on which a computer program is stored, wherein when the computer program is executed by a processor, the steps of the video editing method in the embodiment are implemented.

[0100] In this embodiment, the computer-readable storage medium includes flash memory, hard disk, multimedia card, card-type memory (for example, SD or DX memory, etc.), random access memory (RAM), static random access memory (SRAM), read-only memory (ROM), electrically erasable programmable read-only memory (EEPROM), programmable read-only memory (PROM), magnetic memory, disk, optical disk, etc. In some embodiments, the computer-readable storage medium can be an internal storage unit of a computer device, such as a hard disk or memory of the computer device. In other embodiments, the computer-readable storage medium can also be an external storage device of a computer device, such as a plug-in hard disk equipped on the computer device, a smart memory card (Smart Media Card, SMC), a secure digital (Secure Digital, SD) card, a flash card (Flash Card), etc. Of course, the computer-readable storage medium can also include both the internal storage unit of the computer device and its external storage device. In this embodiment, the computer-readable storage medium is generally used to store an operating system and various application software installed on the computer device, such as the program code of the video editing method in the embodiment. In addition, the computer-readable storage medium can also be used to temporarily store various types of data that have been output or are to be output.

[0101] Embodiment 5 An embodiment of the present application also provides a computer program product, including a computer program, which implements the method in the above embodiment when executed by a processor.

[0102] Obviously, those skilled in the art should understand that the modules or steps of the above-mentioned embodiments of the present application can be implemented by general-purpose computer devices, they can be concentrated on a single computer device, or distributed on a network composed of multiple computer devices, optionally, they can be implemented by executable program codes of computer devices, so that they can be stored in a storage device and executed by the computer device, and in some cases, the steps shown or described can be executed in a different order from that herein, or they can be made into individual integrated circuit modules, or multiple modules or steps therein can be made into a single integrated circuit module for implementation. In this way, the embodiments of the present application are not limited to any specific combination of hardware and software.

[0103] It should be noted that the above are only preferred embodiments of the present application, and the patent protection scope of the present application is not limited thereto. Any equivalent structure or equivalent process transformation made using the contents of the specification and drawings of the present application, or directly or indirectly applied in other related technical fields, are also included in the patent protection scope of the present application.

Claims

1. A video editing method, characterized in that: The method comprises: Get the text description content of the target video; Determining a plurality of candidate segments from the target video according to the text description content; Using a plurality of intelligent agents to score the plurality of candidate segments to obtain scoring results, wherein different intelligent agents give different scores to the candidate segments; Determining a plurality of target segments from the plurality of candidate segments according to the scoring result; The target segment is clipped and the clipped video segment is output.

2. The method according to claim 1, characterized in that The step of determining a plurality of candidate segments from the target video according to the text description content includes: A large language model agent is used to select multiple segments from the target video according to the text description content, and generate structured information for each of the segments to obtain multiple candidate segments containing the structured information, wherein the structured information includes at least part of the segment number, starting position, ending position, segment title and segment content.

3. The method according to claim 2, characterized in that The adopting several intelligent agents to score the multiple candidate segments comprises: Using several intelligent agents to score the structured information of the candidate fragment; and / or obtaining the text fragment corresponding to the candidate fragment from the text description content, and using the several intelligent agents to score the text fragment.

4. The method according to claim 1, characterized in that The step of determining a plurality of candidate segments from the target video according to the text description content includes: Using a topic segmentation agent to determine different topics in the text description content, and dividing the target video into multiple topic segments according to the different topics; A content selection agent is used to determine multiple candidate segments from the multiple topic segments according to the text description content.

5. The method according to claim 1, characterized in that The adopting several intelligent agents to score the multiple candidate segments comprises: Using a content quality assessment agent to score the multiple candidate segments to obtain a content quality score for each candidate segment; Using an attractiveness evaluation agent to score the multiple candidate segments to obtain an attractiveness score for each of the candidate segments; Using a resonance evaluation agent to score the multiple candidate segments to obtain a resonance score for each candidate segment; The multiple candidate segments are scored using a content compliance assessment agent to obtain a content compliance score for each candidate segment.

6. The method according to claim 5, characterized in that Determining a plurality of target segments from the plurality of candidate segments according to the scoring result includes: Using a comprehensive scoring agent to determine a comprehensive score for each candidate segment based on the content quality score, the attractiveness score, the resonance score, and the content compliance score; The comprehensive scoring agent is used to determine the top N candidate segments as the target segments according to the comprehensive scores, where N is a positive integer.

7. The method according to any one of claims 1 to 6, characterized in that: The step of obtaining the text description content of the target video includes: Get the voice content of the target video; Automatic speech recognition is used to convert the speech content into text description content of the target video.

8. A video editing device, characterized in that: The device comprises: An acquisition module is used to obtain the text description content of the target video; A first determination module, configured to determine a plurality of candidate segments from the target video according to the text description content; A scoring module, configured to use a plurality of intelligent agents to score the plurality of candidate segments to obtain scoring results, wherein different intelligent agents give different scores to the candidate segments; A second determination module, configured to determine a plurality of target segments from the plurality of candidate segments according to the scoring result; The editing module is used to edit the target segment and output the edited video segment.

9. A computer device, characterized in that: include: at least one processor; and a memory communicatively connected to the at least one processor; wherein: The memory stores instructions that can be executed by the at least one processor, and the instructions are executed by the at least one processor to enable the at least one processor to perform the method according to any one of claims 1 to 7.

10. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores computer instructions, and when the computer instructions are executed by a processor, the method according to any one of claims 1 to 7 is implemented.

11. A computer program product, comprising a computer program, characterized in that When the computer program is executed by a processor, the steps of the method according to any one of claims 1 to 7 are implemented.

Images