Video pushing method and device, computer equipment and storage medium

Abstract

The embodiment of the invention belongs to the field of image processing, and relates to a video pushing method and device, computer equipment and a storage medium, and the method comprises the following steps: obtaining initial video data; performing feature analysis on the initial video data to obtain video content features; performing content feature analysis on the initial video data based on the video content features to obtain video feature information; obtaining platform feature information, and determining key feature information according to the video feature information and the platform feature information; and performing video reconstruction on the initial video data based on the key feature information to generate effective video data, and pushing the effective video data to a corresponding platform for video playing. According to the method and the device, the video data conforming to different feature platforms can be effectively generated and correspondingly pushed.

Description

Technical Field

[0001] The present application relates to the field of image processing technology, specifically to the field of financial technology, and in particular to a video push method, device, computer equipment and storage medium. Background Art

[0002] With the rapid development of the e-commerce industry, the e-commerce 2.0 era has taken mobile e-commerce and social e-commerce as the main models, greatly improving the convenience of shopping and the diversity of channels. This change has forced companies to carry out refined operations for different channels and quickly expand their businesses to adapt to market changes. However, in this process, the traditional e-commerce sales video production method - manual editing and secondary creation, is not only costly and time-consuming, but also difficult to adapt to the rapid iteration of multiple platforms and the instant changes in user preferences. This challenge becomes particularly arduous, especially when considering the unique user characteristics and content review standards of each platform.

[0003] Moreover, for many small and medium-sized enterprises, it is not easy to use traditional video production methods. This process not only requires professional creative planning and high-quality shooting and production, but also requires a lot of human and material resources, which is undoubtedly a heavy burden for small and medium-sized enterprises with limited resources. In addition, for different video platforms, their audiences are also different. Traditionally shot e-commerce sales videos cannot effectively guarantee their playback effects on different platforms, which not only leads to a waste of human resources, but also fails to play an effective role in product promotion, bringing great difficulties to the company's product promotion work. Summary of the invention

[0004] The purpose of the embodiments of the present application is to propose a video push method, apparatus, computer equipment and storage medium to solve the problem of being unable to effectively generate video data that conforms to different feature platforms and perform corresponding push.

[0005] In order to solve the above technical problems, the embodiment of the present application provides a video push method, which adopts the following technical solution:

[0006] The step of performing feature analysis on the initial video data to obtain video content features specifically includes:

[0007] Dividing the initial video data into segments to obtain video data segments;

[0008] Performing target detection and speech recognition on the video data segment to obtain key frames and speech text;

[0009] Identify the appearance frequency and duration of the main character based on the key frames, and determine the main character of the video according to the appearance frequency and duration;

[0010] Identifying scene type changes in adjacent video data segments, and determining scene switching points according to the scene type changes;

[0011] The key frames, the voice text, the main characters in the video, and the scene switching points are integrated into the video content features.

[0012] Furthermore, the step of performing feature analysis on the initial video data to obtain video content features specifically includes:

[0013] Dividing the initial video data into segments to obtain video data segments;

[0014] Performing target detection and speech recognition on the video data segment to obtain key frames and speech text;

[0015] Identify the appearance frequency and duration of the main character based on the key frames, and determine the main character of the video according to the appearance frequency and duration;

[0016] Identifying scene type changes in adjacent video data segments, and determining scene switching points according to the scene type changes;

[0017] The key frames, the voice text, the main characters in the video, and the scene switching points are integrated into the video content features.

[0018] Furthermore, the step of performing content feature analysis on the initial video data based on the video content feature to obtain video feature information specifically includes:

[0019] Extracting features from the initial video data based on the video content features to obtain visual features, audio features, and text features;

[0020] Constructing a content feature vector according to the visual feature, the audio feature, and the text feature;

[0021] A preset classification prediction model is obtained, and the content feature vector is input into the preset classification prediction model to perform content feature analysis to obtain the video feature information.

[0022] Furthermore, the step of obtaining platform characteristic information and determining key characteristic information according to the video characteristic information and the platform characteristic information specifically includes:

[0023] Acquire platform communication characteristic data, and perform feature extraction based on the platform communication characteristic data to obtain the platform characteristic information;

[0024] Matching the video feature information with the platform feature information to obtain platform matching information;

[0025] The video feature information is screened according to the platform matching information to obtain the key feature information.

[0026] Furthermore, the step of reconstructing the initial video data based on the key feature information to generate valid video data specifically includes:

[0027] Extract key information from the initial video data according to the key feature information to obtain key segments and important elements;

[0028] Reconstruct the video based on the genetic algorithm, the key segments, and the important elements to generate a reconstructed video;

[0029] Obtaining platform requirement information, and determining whether the video content of the reconstructed video meets the platform requirement information;

[0030] If the video content of the reconstructed video meets the platform requirement information, the reconstructed video is used as the valid video data.

[0031] Furthermore, the step of reconstructing the video based on the genetic algorithm, the key segments, and the important elements to generate the reconstructed video specifically includes:

[0032] Acquire platform videos, and cluster them according to the playback performance of the platform videos to obtain performance video classes;

[0033] Calculate the comprehensive performance score of the performance video category, and perform matching and screening in the performance video category according to the key segments, the important elements, and the comprehensive performance score to obtain the optimal platform video;

[0034] The performance features of the platform video are extracted, and based on a genetic algorithm, the performance features of the platform video are used as an initial population, and the optimal platform video is used as a reference to perform video reconstruction to generate the reconstructed video.

[0035] Furthermore, the step of acquiring platform videos and clustering them according to the playback performance of the platform videos to obtain the performance video class specifically includes:

[0036] Get the playback performance data of platform videos;

[0037] Extracting key elements from the platform video and establishing a video feature vector based on the key elements;

[0038] The platform videos are divided into different performance video classes according to a clustering algorithm and the video feature vector.

[0039] In order to solve the above technical problems, the embodiment of the present application further provides a video push device, which adopts the following technical solution:

[0040] A data acquisition module, used for acquiring initial video data;

[0041] A feature analysis module, used to perform feature analysis on the initial video data to obtain video content features;

[0042] A feature analysis module, configured to perform content feature analysis on the initial video data based on the video content features to obtain video feature information;

[0043] An attribute acquisition module, used to acquire platform characteristic information, and determine key characteristic information according to the video characteristic information and the platform characteristic information;

[0044] The video push module is used to reconstruct the initial video data based on the key feature information, generate valid video data, and push the valid video data to the corresponding platform for video playback.

[0045] In order to solve the above technical problems, the embodiment of the present application further provides a computer device, which adopts the following technical solution:

[0046] A computer device comprises a memory and a processor, wherein the memory stores computer-readable instructions, and the processor implements the steps of the video pushing method as described in any one of the above items when executing the computer-readable instructions.

[0047] In order to solve the above technical problems, the embodiment of the present application further provides a computer-readable storage medium, which adopts the following technical solution:

[0048] A computer-readable storage medium stores computer-readable instructions, and when the computer-readable instructions are executed by a processor, the steps of the video pushing method as described in any one of the above items are implemented.

[0049] Compared with the prior art, the embodiment of the present application has the following beneficial effects: the embodiment obtains initial video data; performs feature analysis on the initial video data to obtain video content features; performs content feature analysis on the initial video data based on the video content features to obtain video feature information; obtains platform feature information, determines key feature information based on the video feature information and the platform feature information; performs video reconstruction on the initial video data based on the key feature information to generate valid video data, and pushes the valid video data to the corresponding platform for video playback. This effectively generates video data that conforms to different feature platforms and pushes them accordingly. BRIEF DESCRIPTION OF THE DRAWINGS

[0050] In order to more clearly illustrate the scheme in the present application, a brief introduction is given below to the drawings required for use in the description of the embodiments of the present application. Obviously, the drawings described below are some embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative work.

[0051] Figure 1 is an exemplary system architecture diagram to which the present application may be applied;

[0052] Figure 2 A flowchart of an embodiment of a video push method according to the present application;

[0053] Figure 3 yes Figure 2 A flowchart of a specific implementation of step S20;

[0054] Figure 4 yes Figure 2 A flowchart of a specific implementation of step S30;

[0055] Figure 5 yes Figure 2 A flowchart of a specific implementation of step S40;

[0056] Figure 6 yes Figure 2 A flowchart of a specific implementation of step S50;

[0057] Figure 7 yes Figure 6 A flowchart of a specific implementation of step S502;

[0058] Figure 8 yes Figure 7 A flowchart of a specific implementation of step S5021;

[0059] Fig. 9 is a structural schematic diagram of an embodiment of a video push device according to the present application;

[0060] Fig.10 It is a structural diagram of an embodiment of a computer device according to the present application. DETAILED DESCRIPTION

[0061] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as those commonly understood by technicians in the technical field of the present application; the terms used in the specification of the application herein are only for the purpose of describing specific embodiments and are not intended to limit the present application; the terms "including" and "having" and any variations thereof in the specification and claims of the present application and the above-mentioned drawings are intended to cover non-exclusive inclusions. The terms "first", "second", etc. in the specification and claims of the present application or the above-mentioned drawings are used to distinguish different objects, not to describe a specific order.

[0062] Reference to "embodiments" herein means that a particular feature, structure, or characteristic described in conjunction with the embodiments may be included in at least one embodiment of the present application. The appearance of the phrase in various locations in the specification does not necessarily refer to the same embodiment, nor does it refer to non-related or alternative embodiments that are mutually exclusive with other embodiments. It is explicitly and implicitly understood by those skilled in the art that the embodiments described herein may be combined with other embodiments.

[0063] In order to enable those skilled in the art to better understand the solution of the present application, the technical solution in the embodiments of the present application will be clearly and completely described below in conjunction with the accompanying drawings.

[0064] like Figure 1 As shown, the system architecture 100 may include terminal devices 101, 102, 103, a network 104 and a server 105. The network 104 is used to provide a medium for communication links between the terminal devices 101, 102, 103 and the server 105. The network 104 may include various connection types, such as wired, wireless communication links or optical fiber cables, etc.

[0065] Users can use terminal devices 101, 102, 103 to interact with server 105 through network 104 to receive or send messages, etc. Various communication client applications can be installed on terminal devices 101, 102, 103, such as web browser applications, shopping applications, search applications, instant messaging tools, email clients, social platform software, etc.

[0066] Terminal devices 101, 102, 103 can be various electronic devices with display screens and supporting web browsing, including but not limited to smart phones, tablet computers, e-book readers, MP3 players (Moving Picture Experts Group Audio Layer III), MP4 (Moving Picture Experts Group Audio Layer IV) players, laptop computers, desktop computers, etc.

[0067] The server 105 may be a server that provides various services, such as a background server that provides support for web pages displayed on the terminal devices 101 , 102 , and 103 .

[0068] It should be noted that the video push method provided in the embodiment of the present application is generally executed by a server, and accordingly, the video push device is generally arranged in the server.

[0069] It should be understood that Figure 1 The number of terminal devices, networks and servers in the embodiment is only for illustration. Any number of terminal devices, networks and servers may be provided according to implementation requirements.

[0070] Continue to refer Figure 2 , shows a flow chart of an embodiment of a method for video push according to the present application. The video push method comprises the following steps:

[0071] Step S10, obtaining initial video data;

[0072] In this embodiment, the initial video data refers to the video directly obtained from the data source (such as a website, a system video library, etc.). The initial video data can be a product sales video or a product promotion video. The product sales video increases the customer's desire to buy by displaying basic information such as the appearance, color, and texture of the product, as well as functional information such as the use process and effect demonstration. The product promotion video displays the product in real time in the form of live broadcast and interacts with the audience, thereby promoting the sale of the product. In this embodiment, the initial video data is stored in the database, and a matching query can be performed in the database according to the video data extraction identifier, so as to effectively obtain the initial video data.

[0073] Step S20, performing feature analysis on the initial video data to obtain video content features;

[0074] In this embodiment, the video content features include key frames of the initial video data, scene switching points, main characters in the video, voice text and other information. Key frames are frames in the video that can represent its important content or change points. In the commodity sales video, key frames can include close-up shots of the commodity, key steps in the use demonstration, and the highlights of the anchor's explanation. The scene switching point refers to the moment when the video switches from one scene to another. In the commodity sales video, the scene switching occurs when showing different commodities, changing the shooting environment, or switching the explanation theme. The main characters in the video refer to the characters who appear frequently in the video and have an important influence on the video content. In the commodity sales video, the main characters are usually the anchor or the product spokesperson. Voice text is the text information generated by converting the voice content in the video. In the commodity sales video, the voice text contains the anchor's explanation content, product introduction, promotion information and other key information.

[0075] Step S30, performing content feature analysis on the initial video data based on the video content feature to obtain video feature information;

[0076] In this embodiment, the video feature information includes the video style, video theme, video audience attributes, etc. corresponding to the video content of the initial video data. The video style refers to the overall artistic characteristics and forms of expression, including visual effects: such as color matching, filter effects, animation application, etc., editing rhythm: the speed of the video, the frequency of camera switching, etc., background music and sound effects: what kind of music is selected as background music, and the use of sound effects. The video theme refers to the central idea or main content around which the video revolves. Taking the commodity sales video as an example, the video theme clearly points to a certain commodity type, functional characteristics, usage scenario or promotional activity. For example, the sales video of kitchen utensils may be themed "efficient and convenient cooking experience". The video audience attributes refer to the characteristics of the expected target audience of the video, including age group: people of different age groups have different preferences and acceptance of video content, gender: gender differences lead to different preferences for video style and theme, hobbies: the theme and content of the video are related to specific hobbies, attracting audiences with corresponding hobbies, and purchase intention: the audience's purchase intention determines the degree of attraction of the video to the audience.

[0077] Step S40, obtaining platform characteristic information, and determining key characteristic information according to the video characteristic information and the platform characteristic information;

[0078] In this embodiment, the key feature information is the video feature information after information sorting. Platform feature information refers to the features extracted from the video dissemination feature data of different platforms. The video dissemination feature data includes platform user attributes, platform video content information, interactive behavior information, etc. Platform user attributes include user portraits: basic information such as user age distribution, gender ratio, geographical distribution, and educational level, interest preferences: interest preferences reflected by analyzing users' browsing, liking, commenting, sharing and other behaviors on the platform, and consumption capacity: consumption capacity reflected by users' willingness to buy and consumption habits. Interactive behavior information includes interaction frequency: the number of times users log in to the platform, the length of time they watch videos, the frequency of interactive behaviors such as likes, comments, and sharing, and interaction depth: the degree of positivity of user interactive behaviors such as likes, comments, and sharing.

[0079] Step S50, reconstructing the initial video data based on the key feature information to generate valid video data, and pushing the valid video data to a corresponding platform for video playback.

[0080] In this embodiment, video reconstruction refers to the process of regenerating a video based on the initial video data and key feature information, and the effective video data is the video data regenerated after the video reconstruction. The platform refers to an online service or application for publishing, disseminating and displaying video content. In this embodiment, the effective video data includes multiple ones, which correspond to the platform requirements of different platforms. The video is pushed by pushing the effective video data to the corresponding platform for video playback.

[0081] This embodiment obtains initial video data; performs feature analysis on the initial video data to obtain video content features; performs content feature analysis on the initial video data based on the video content features to obtain video feature information; obtains platform feature information, determines key feature information based on the video feature information and the platform feature information; performs video reconstruction on the initial video data based on the key feature information to generate valid video data, and pushes the valid video data to the corresponding platform for video playback. This effectively generates video data that conforms to different feature platforms and pushes them accordingly.

[0082] Continue to refer Figure 3 In some optional implementations of this embodiment, step S20 includes the following steps:

[0083] Step S201, dividing the initial video data into segments to obtain video data segments;

[0084] In this embodiment, the video data segment refers to a partial video segment of the initial video data. The video data segment can evenly divide the initial video data according to time intervals (such as every 5 minutes, 10 minutes, etc.), or intelligently divide according to natural breakpoints of the video content (such as advertising insertion, chapter transitions, etc.), or use video processing technology to identify silent periods, black fields, etc. in the video as the basis for division.

[0085] Step S202, performing target detection and speech recognition on the video data segment to obtain key frames and speech text;

[0086] In this embodiment, target detection is to perform image analysis on each video data segment to identify key targets in the video data segment. In this embodiment, a deep learning model (such as YOLO, SSD, etc.) can be used to identify key target objects, especially people, in the video data segment screen, so as to effectively identify key frames. Speech recognition is a method of transcribing the audio in the video data segment and converting speech into text. In this embodiment, automatic speech recognition (ASR) technology can be used to perform speech recognition operations.

[0087] Step S203, identifying the appearance frequency and duration of the main character based on the key frames, and determining the main character of the video according to the appearance frequency and duration;

[0088] In this embodiment, by further analyzing the identified key frames, the appearance frequency and duration of the main characters are counted, and the one with the most comprehensive consideration of appearance frequency and duration is taken as the main character of the video. In this embodiment, the same character in different key frames can be confirmed by face recognition technology.

[0089] Step S204, identifying scene type changes in adjacent video data segments, and determining scene switching points according to the scene type changes;

[0090] In this embodiment, by analyzing the image content in adjacent video data segments, the scene type (such as indoor, outdoor, night scene, day scene, etc.) is identified, so as to determine the scene switching point during the scene change process, wherein the scene switching point can be accompanied by visual clues such as lighting changes, background changes, and people entering and exiting.

[0091] Step S205, integrating the key frame, the voice text, the main characters in the video, and the scene switching points into the video content features.

[0092] In this embodiment, by integrating key frames, voice text, main characters in the video, scene switching points and other information together, a comprehensive description of the video content characteristics is formed. In this embodiment, structured data (such as JSON, XML) can be used to organize this information to facilitate subsequent processing and analysis.

[0093] This embodiment divides the initial video data into segments to obtain video data segments; performs target detection and voice recognition on the video data segments to obtain key frames and voice text; identifies the appearance frequency and duration of the main characters based on the key frames, and determines the main characters of the video according to the appearance frequency and duration; identifies the scene type changes in adjacent video data segments, and determines the scene switching points according to the scene type changes; and integrates the key frames, the voice text, the main characters of the video, and the scene switching points into the video content features. Thus, the video content features including key frames, voice text, main characters of the video, scene switching points and other information are effectively extracted from the initial video data to facilitate the subsequent content feature analysis and processing.

[0094] Continue to refer Figure 4 In some optional implementations of this embodiment, step S30 includes the following steps:

[0095] Step S301, extracting features from the initial video data based on the video content features to obtain visual features, audio features, and text features;

[0096] In this embodiment, visual features include color features, texture features, shape features, motion features, scene features, etc., audio features include voice features, background music features, sound effect features, etc., and text features include bag-of-words models, TF-IDF vectors, topic models, text sentiment, etc. In this embodiment, feature extraction can be performed using corresponding image processing tools, audio processing tools, and text processing tools.

[0097] Step S302, constructing a content feature vector according to the visual feature, the audio feature, and the text feature;

[0098] In this embodiment, the visual features, audio features, and text features are encoded to obtain corresponding visual feature codes, audio feature codes, and text feature codes. The visual feature code can be obtained by splicing the extracted feature vectors of color, texture, shape, etc. The audio feature code can be obtained by splicing or converting the extracted audio feature vectors such as MFCC and fundamental frequency into a fixed-length representation. The text feature code is represented as a bag-of-words model, a TF-IDF vector, or a topic model distribution. The visual feature code, the audio feature code, and the text feature code are weightedly spliced ​​to obtain a content feature vector.

[0099] Step S303, obtaining a preset classification prediction model, inputting the content feature vector into the preset classification prediction model to perform content feature analysis, and obtaining the video feature information.

[0100] In this embodiment, the video feature information includes the video style, video theme, and video audience attributes corresponding to the video content of the initial video data. The preset classification prediction model includes a video style classification model, a video theme classification model, and an audience attribute prediction model. Among them, the video style classification model uses a support vector machine algorithm to classify video styles, the video theme classification model uses a convolutional neural network algorithm to classify video themes, and the audience attribute prediction model uses a logistic regression algorithm to predict video audience attributes.

[0101] This embodiment extracts features from the initial video data based on the video content features to obtain visual features, audio features, and text features; constructs a content feature vector based on the visual features, the audio features, and the text features; obtains a preset classification prediction model, and inputs the content feature vector into the preset classification prediction model for content feature analysis, thereby effectively obtaining video feature information about the visual features, audio features, and text features of the initial video data, so as to facilitate subsequent determination of key feature information.

[0102] Continue to refer Figure 5 In some optional implementations of this embodiment, step S40 includes the following steps:

[0103] Step S401, obtaining platform communication characteristic data, and performing feature extraction based on the platform communication characteristic data to obtain the platform characteristic information;

[0104] In this embodiment, the platform communication characteristic data includes platform user attributes, platform video content information, interactive behavior information, etc., and the platform characteristic information includes user attribute characteristics, platform video content characteristics, and interactive behavior characteristics. Among them, user attribute characteristics include: basic information, interest preferences, and user activity; platform video content characteristics include content type, content quality, and content trends; interactive behavior characteristics include: user interactive behavior, social relationships, and participation.

[0105] Step S402, matching the video feature information with the platform feature information to obtain platform matching information;

[0106] In this embodiment, the video feature information and the platform feature information are matched by similarity calculation. In this embodiment, the similarity calculation adopts cosine similarity, by quantifying the video style, video theme, and video audience attributes in the video feature information, for example, quantifying the video style into a set of style labels, each label corresponds to a specific style dimension (such as comedy, suspense, science fiction, etc.), using 0 / 1 coding or probability value to indicate whether the video belongs to or tends to a certain style, and then constructing corresponding feature vectors for the quantified video feature information and platform feature information, and calculating the similarity between the feature vector of the video feature information and the feature vector constructed by the platform feature information by cosine similarity, which is an indicator to measure the degree of similarity between two vectors in direction, and its value ranges from -1 to 1. The calculated cosine similarity score is used as the platform matching information that characterizes the matching relationship between the video feature and the platform feature.

[0107] Step S403: Filter the video feature information according to the platform matching information to obtain the key feature information.

[0108] In this embodiment, the video style, video theme, and video audience attribute respectively correspond to different platform matching degrees. The video feature information is screened by the cosine similarity score corresponding to the platform matching information. The video feature information corresponds to a matching threshold. The cosine similarity score corresponding to the platform matching information is compared with the matching threshold. The video feature information (video style, video theme, video audience attribute) with a cosine similarity score greater than or equal to the matching threshold is used as the key feature information.

[0109] This embodiment obtains the platform propagation characteristic data and performs feature extraction based on the platform propagation characteristic data to obtain the platform characteristic information; matches the video characteristic information with the platform characteristic information to obtain platform matching information; and filters the video characteristic information according to the platform matching information to obtain key feature information in the video characteristic information that is closely related to the platform characteristics, so as to facilitate subsequent video reconstruction processing.

[0110] Continue to refer Figure 6 In some optional implementations of this embodiment, step S50 includes the following steps:

[0111] Step S501, extracting key information from the initial video data according to the key feature information to obtain key segments and important elements;

[0112] In this embodiment, key information extraction includes key segment extraction and important element recognition. Key segment extraction refers to the use of video analysis technology (such as scene switching detection, face recognition, action recognition, etc.) to identify key scenes or segments related to key feature information in the initial video data. Important element recognition refers to the use of video analysis technology (such as text image recognition, voice recognition, etc.) to identify key objects, text, sounds and other elements related to key feature information in the initial video data.

[0113] Step S502, reconstructing the video based on the genetic algorithm, the key segments, and the important elements to generate a reconstructed video;

[0114] In this embodiment, the genetic algorithm includes parameter encoding: converting the solution space of the problem into a search space that can be processed by the genetic algorithm. Common encoding methods include binary encoding, floating-point encoding, symbolic encoding, etc.; setting of the initial population: randomly generating a certain number of individuals as the initial population, and these individuals represent potential solutions to the problem; design of the fitness function: used to evaluate the fitness of individuals in the population, that is, the degree of adaptation of the individuals to the environment, usually designed according to the objective function of the problem; genetic operation design: including operations such as selection, crossover and mutation, which simulate the genetic mechanism in the process of biological evolution and are used to generate new populations; control parameter setting: including population size, crossover probability, mutation probability, etc., these parameters have a great influence on the performance of the genetic algorithm.

[0115] Step S503, obtaining platform requirement information, and determining whether the video content of the reconstructed video meets the platform requirement information;

[0116] In this embodiment, the platform requirement information refers to a series of regulations and restrictions set by the platform for playing video content. The platform requirement information includes content theme: confirming whether the theme of the reconstructed video is consistent with the platform's positioning and user preferences; video length: checking whether the video length is within the range specified by the platform; video quality: ensuring that parameters such as video resolution and frame rate meet the platform requirements; copyright ownership: verifying the copyright status of the video content to ensure that no copyright is infringed; prohibited content: carefully checking the video content to ensure that it does not contain any elements prohibited by the platform. By comparing the content theme, video length, video quality, copyright ownership, and prohibited content of the reconstructed video with the platform requirement information item by item, it is determined whether the video content of the reconstructed video meets the platform requirement information.

[0117] Step S504: if the video content of the reconstructed video meets the platform requirement information, the reconstructed video is used as the valid video data;

[0118] In this embodiment, the video content of the reconstructed video meets the platform requirement information, which means that the content theme, video length, video quality, copyright ownership, and prohibited content of the reconstructed video all correspond to the platform requirement information. Only when all of them correspond, the reconstructed video can be used as the valid video data.

[0119] Step S505: If the video content of the reconstructed video does not meet the platform requirement information, the reconstructed video is corrected according to the platform requirement information, and the corrected reconstructed video is used as the valid video data.

[0120] In this embodiment, correcting the reconstructed video according to the platform requirement information refers to modifying one or more of the content theme, video length, video quality, copyright ownership, and prohibited content of the reconstructed video to comply with the platform requirement information, for example, modifying the content theme of the reconstructed video to the platform positioning and user preferences in the platform requirement information.

[0121] This embodiment extracts key information from the initial video data according to the key feature information to obtain key segments and important elements; reconstructs the video based on the genetic algorithm, the key segments, and the important elements to generate a reconstructed video; obtains platform requirement information to determine whether the video content of the reconstructed video meets the platform requirement information; if the video content of the reconstructed video meets the platform requirement information, the reconstructed video is used as the valid video data. In this way, valid video data that takes into account the key segments and important elements of the initial video data and meets the platform requirement information corresponding to the platform is effectively generated, so as to facilitate the subsequent push of valid video data on the platform.

[0122] Continue to refer Figure 7In some optional implementations of this embodiment, step S502 includes the following steps:

[0123] Step S5021, obtaining platform videos, and clustering them according to the playback performance of the platform videos to obtain performance video classes;

[0124] In this embodiment, platform videos are collected and acquired by the corresponding platform. The playback performance of platform videos refers to performance data such as the volume of views, likes, comments, and reposts of the platform videos on the platform. Clustering is performed based on the playback performance of the platform videos, so that platform videos with similar volume of views, likes, comments, and reposts are clustered into performance video categories.

[0125] Step S5022, calculating the comprehensive performance score of the performance video category, performing matching and screening in the performance video category according to the key segments, the important elements, and the comprehensive performance score, to obtain the optimal platform video;

[0126] In this embodiment, the comprehensive performance score of the performance video category is obtained by weighted calculation, wherein each playback performance (playback volume, number of likes, number of comments, number of reposts, etc.) in the performance video category is set with a different weight, which can be preset and adjusted according to the actual situation. In this embodiment, the weights of the playback volume, number of likes, number of comments, and number of reposts are all set to 0.25. The playback performance data of the playback volume, number of likes, number of comments, and number of reposts are indexed, and then weighted calculation is performed according to the indexed index value and the corresponding weight, and the weighted calculated value is summed to obtain the comprehensive performance score corresponding to the performance video category. According to the similarity between the key segments, important elements and the performance video category, and considering the comprehensive performance score, matching and screening are performed in the performance video category to obtain the optimal platform video considering the comprehensive key segments, important elements, and comprehensive performance score.

[0127] Step S5023, extracting the performance features of the platform video, using the performance features of the platform video as an initial population based on a genetic algorithm, and using the optimal platform video as a reference to perform video reconstruction to generate the reconstructed video.

[0128] In this embodiment, when the performance characteristics of the platform video are used as the initial population and the optimal platform video is used as a reference, the initial population is reconstructed according to selection, crossover, and mutation operations. The selection operation is to select a certain number of individuals from the current population as parents based on the result of the fitness function to generate the next generation. The selection operation should tend to select individuals with high fitness. The crossover operation is to cross (or recombine) the selected parent individuals to generate new individuals. The crossover operation simulates the gene recombination phenomenon in the process of biological evolution, aiming to combine the advantages of parent individuals. The mutation operation is to mutate the new individuals generated after the crossover with a certain probability, that is, to randomly change the values ​​of certain features in the genetic code of the individual. The mutation operation increases the diversity of the population and helps to avoid falling into the local optimal solution.

[0129] This embodiment obtains platform videos, clusters them according to their playback performance, and obtains performance video classes; calculates the comprehensive performance scores of the performance video classes, performs matching and screening in the performance video classes according to the key segments, the important elements, and the comprehensive performance scores, and obtains the optimal platform video; extracts the performance features of the platform video, and reconstructs the video based on a genetic algorithm using the performance features of the platform video as the initial population and the optimal platform video as a reference, thereby effectively obtaining a reconstructed video that integrates the key segments, important elements, and platform videos, so as to facilitate subsequent platform requirement information compliance judgment processing.

[0130] Continue to refer Figure 8 In some optional implementations of this embodiment, step S5021 includes the following steps:

[0131] Step S50211, obtaining playback performance data of platform videos;

[0132] In this embodiment, the playback performance data of the platform video includes data such as the number of playbacks, the number of likes, the number of comments, and the number of reposts, which can be extracted from the platform database through a performance data extraction identifier or collected and obtained by the corresponding platform.

[0133] Step S50212, extracting key elements from the platform video, and establishing a video feature vector based on the key elements;

[0134] In this embodiment, the key elements include video duration, picture style, music rhythm and subtitle position. The extraction of key elements includes extracting video duration: directly reading from the metadata or attributes of the video file, or obtaining the duration by playing the video and calculating the playing time; extracting picture style: visual feature analysis: using computer vision technology to analyze the frames of the video, extracting features such as color distribution, texture, brightness contrast, etc., scene recognition: further analyzing the scene changes in the video, and identifying different scene types (such as indoor, outdoor, natural scenery, urban landscape, etc.); extracting music rhythm: audio analysis: using audio analysis technology to extract audio signals in the video, and analyze their rhythm, pitch, tone, etc. Features such as quantity, rhythm detection: identify the rhythm of the soundtrack by detecting the beat and rhythm pattern in the audio signal, feature quantization: quantify the rhythm features into numerical values ​​for subsequent processing and analysis; extract subtitle position: OCR technology: use optical character recognition (OCR) technology to identify subtitle text embedded in the video frame in text form, subtitle position detection: after identifying the subtitle text, further analyze the position of the subtitle in the video frame, which can be achieved through image processing and computer vision technology, such as edge detection, contour recognition, etc., position coordinate extraction: quantify the position of the subtitle into specific coordinate values ​​(such as pixel coordinates or proportional coordinates relative to the video frame) for subsequent processing and analysis. Feature vector construction is achieved by converting key elements into numerical features and combining them. The numerical feature conversion of video duration includes: normalizing the duration to a certain range, such as between 0 and 1; the numerical feature conversion of picture style includes: using One-Hot Encoding or Label Encoding to convert the picture style of each video into a numerical vector; the numerical feature conversion of music rhythm includes: calculating the beats per minute (BPM) to quantify and use it as an element in the feature vector; the numerical feature conversion of subtitle position includes: quantifying the subtitle position coordinates in the video frame (such as the upper left corner, lower right corner, etc.) and converting them into numerical representation. If the subtitle position is dynamic, it can be quantified by calculating its average position or the most frequently occurring position. After obtaining the above quantized key elements, the video feature vector is obtained by combining the quantized key elements, for example, video duration (normalized): 1 element, picture style (one-hot encoding): assuming there are 3 styles, then there are 3 elements (0, 1, 0 represents the second style), music rhythm (BPM + rhythm change rate): 2 elements, subtitle position (coordinate quantization + proportion): assuming the coordinates of the upper left corner and the lower right corner, and the proportion of subtitles are used, then there are 4 elements. Therefore, a video feature vector containing all the above features will have 1 (duration) + 3 (style) + 2 (rhythm) + 4 (subtitle position) = 10 elements.

[0135] Step S50213, dividing the platform videos into different performance video classes according to the clustering algorithm and the video feature vector.

[0136] In this embodiment, the clustering algorithm can adopt K-means. By configuring the number of clusters K in K-means, the video feature vector is used as the algorithm input, and then the K-means clustering algorithm is executed to obtain the clustering results. The platform videos are divided according to the clustering results to obtain the performance video class.

[0137] This embodiment obtains the playback performance data of the platform video; extracts key elements from the platform video and establishes a video feature vector based on the key elements; and divides the platform video into different performance video classes based on the clustering algorithm and the video feature vector. This effectively achieves accurate classification of the performance video classes, facilitating the subsequent calculation of the comprehensive performance score.

[0138] Those skilled in the art can understand that all or part of the processes in the above-mentioned embodiments can be implemented by instructing the relevant hardware through computer-readable instructions, and the computer-readable instructions can be stored in a computer-readable storage medium. When the program is executed, it can include the processes of the embodiments of the above-mentioned methods. Among them, the aforementioned storage medium can be a non-volatile storage medium such as a disk, an optical disk, a read-only memory (ROM), or a random access memory (RAM).

[0139] It should be understood that, although the steps in the flowchart of the accompanying drawings are displayed in sequence as indicated by the arrows, these steps are not necessarily executed in sequence in the order indicated by the arrows. Unless otherwise specified herein, there is no strict order restriction on the execution of these steps, and they can be executed in other orders. Moreover, at least a part of the steps in the flowchart of the accompanying drawings may include multiple sub-steps or multiple stages, and these sub-steps or stages are not necessarily executed at the same time, but can be executed at different times, and their execution order is not necessarily sequential, but can be executed in turn or alternately with other steps or at least a part of the sub-steps or stages of other steps.

[0140] Further references Fig. 9 , as a response to the above Figure 1 The present application provides an embodiment of a video push device, and the device embodiment is similar to Figure 1 Corresponding to the method embodiment shown, the device can be specifically applied to various electronic devices.

[0141] like Fig. 9As shown, the video push device 600 described in this embodiment includes: a data acquisition module 601, a feature analysis module 602, a characteristic analysis module 603, an attribute acquisition module 604, and a video push module 605. Among them:

[0142] The data acquisition module 601 is used to acquire initial video data;

[0143] A feature analysis module 602 is used to perform feature analysis on the initial video data to obtain video content features;

[0144] A feature analysis module 603 is used to perform content feature analysis on the initial video data based on the video content feature to obtain video feature information;

[0145] The attribute acquisition module 604 is used to acquire platform characteristic information and determine key characteristic information according to the video characteristic information and the platform characteristic information;

[0146] The video push module 605 is used to reconstruct the initial video data based on the key feature information, generate valid video data, and push the valid video data to the corresponding platform for video playback.

[0147] By adopting the above-mentioned video push device, this embodiment can obtain initial video data; perform feature analysis on the initial video data to obtain video content features; perform content feature analysis on the initial video data based on the video content features to obtain video feature information; obtain platform feature information, determine key feature information based on the video feature information and the platform feature information; perform video reconstruction on the initial video data based on the key feature information to generate valid video data, and push the valid video data to the corresponding platform for video playback. In this way, video data that meets different feature platforms can be effectively generated and pushed accordingly.

[0148] To solve the above technical problems, the present application also provides a computer device. Fig.10 , Fig.10 This is a basic structural block diagram of the computer device in this embodiment.

[0149] The computer device 7 includes a memory 71, a processor 72, and a network interface 73 that are interconnected through a system bus. It should be noted that the figure only shows a computer device 7 with components 71-73, but it should be understood that it is not required to implement all the components shown, and more or fewer components can be implemented instead. Among them, those skilled in the art can understand that the computer device here is a device that can automatically perform numerical calculations and / or information processing according to pre-set or stored instructions, and its hardware includes but is not limited to microprocessors, application specific integrated circuits (Application Specific Integrated Circuit, ASIC), programmable gate arrays (Field-Programmable Gate Array, FPGA), digital processors (Digital Signal Processor, DSP), embedded devices, etc.

[0150] The computer device may be a computing device such as a desktop computer, a notebook, a PDA, a cloud server, etc. The computer device may interact with a user through a keyboard, a mouse, a remote controller, a touch pad, or a voice control device.

[0151] The memory 71 includes at least one type of readable storage medium, and the 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 memory 71 can be an internal storage unit of the computer device 7, such as a hard disk or memory of the computer device 7. In other embodiments, the memory 71 can also be an external storage device of the computer device 7, such as a plug-in hard disk equipped on the computer device 7, a smart memory card (Smart Media Card, SMC), a secure digital (Secure Digital, SD) card, a flash card (FlashCard), etc. Of course, the memory 71 can also include both the internal storage unit of the computer device 7 and its external storage device. In this embodiment, the memory 71 is generally used to store the operating system and various application software installed on the computer device 7, such as computer-readable instructions of the video push method, etc. In addition, the memory 71 can also be used to temporarily store various types of data that have been output or are to be output.

[0152] The processor 72 may be a central processing unit (CPU), a controller, a microcontroller, a microprocessor, or other data processing chip in some embodiments. The processor 72 is generally used to control the overall operation of the computer device 7. In this embodiment, the processor 72 is used to run computer-readable instructions stored in the memory 71 or process data, such as computer-readable instructions for running the video push method.

[0153] The network interface 73 may include a wireless network interface or a wired network interface. The network interface 73 is generally used to establish a communication connection between the computer device 7 and other electronic devices.

[0154] By using the above-mentioned computer device, this embodiment can obtain initial video data; perform feature analysis on the initial video data to obtain video content features; perform content feature analysis on the initial video data based on the video content features to obtain video feature information; obtain platform feature information, determine key feature information based on the video feature information and the platform feature information; perform video reconstruction on the initial video data based on the key feature information to generate valid video data, and push the valid video data to the corresponding platform for video playback. This effectively generates video data that conforms to different feature platforms and pushes them accordingly.

[0155] The present application also provides another implementation, namely, providing a computer-readable storage medium, wherein the computer-readable storage medium stores computer-readable instructions, and the computer-readable instructions can be executed by at least one processor to enable the at least one processor to perform the steps of the video push method as described above.

[0156] By adopting the above-mentioned computer-readable storage medium, this embodiment can obtain initial video data; perform feature analysis on the initial video data to obtain video content features; perform content feature analysis on the initial video data based on the video content features to obtain video feature information; obtain platform feature information, determine key feature information based on the video feature information and the platform feature information; perform video reconstruction on the initial video data based on the key feature information to generate valid video data, and push the valid video data to the corresponding platform for video playback. In this way, video data that meets different feature platforms can be effectively generated and pushed accordingly.

[0157] Through the description of the above implementation methods, those skilled in the art can clearly understand that the above-mentioned embodiment methods can be implemented by means of software plus a necessary general hardware platform, and of course by hardware, but in many cases the former is a better implementation method. Based on such an understanding, the technical solution of the present application, or the part that contributes to the prior art, can be embodied in the form of a software product, which is stored in a storage medium (such as ROM / RAM, magnetic disk, optical disk), and includes a number of instructions for a terminal device (which can be a mobile phone, computer, server, air conditioner, or network device, etc.) to execute the methods described in each embodiment of the present application.

[0158] Obviously, the embodiments described above are only some embodiments of the present application, rather than all embodiments. The preferred embodiments of the present application are given in the accompanying drawings, but they do not limit the patent scope of the present application. The present application can be implemented in many different forms. On the contrary, the purpose of providing these embodiments is to make the understanding of the disclosure of the present application more thorough and comprehensive. Although the present application is described in detail with reference to the aforementioned embodiments, for those skilled in the art, it is still possible to modify the technical solutions recorded in the aforementioned specific implementation methods, or to perform equivalent replacement of some of the technical features therein. Any equivalent structure made using the contents of the specification and drawings of this application, directly or indirectly used in other related technical fields, is similarly within the scope of patent protection of this application.

[0159] The non-Company software tools or components appearing in the embodiments of this application are merely examples and do not represent actual use.

Claims

1. A video push method, characterized in that: The steps include: Get initial video data; Performing feature analysis on the initial video data to obtain video content features; Performing content feature analysis on the initial video data based on the video content feature to obtain video feature information; Acquire platform characteristic information, and determine key characteristic information according to the video characteristic information and the platform characteristic information; The initial video data is reconstructed based on the key feature information to generate valid video data, and the valid video data is pushed to a corresponding platform for video playback.

2. The video push method according to claim 1, characterized in that: The step of performing feature analysis on the initial video data to obtain video content features specifically includes: Dividing the initial video data into segments to obtain video data segments; Performing target detection and speech recognition on the video data segment to obtain key frames and speech text; Identify the appearance frequency and duration of the main character based on the key frames, and determine the main character of the video according to the appearance frequency and duration; Identifying scene type changes in adjacent video data segments, and determining scene switching points according to the scene type changes; The key frames, the voice text, the main characters in the video, and the scene switching points are integrated into the video content features.

3. The video push method according to claim 1, characterized in that: The step of performing content feature analysis on the initial video data based on the video content feature to obtain video feature information specifically includes: Extracting features from the initial video data based on the video content features to obtain visual features, audio features, and text features; Constructing a content feature vector according to the visual feature, the audio feature, and the text feature; A preset classification prediction model is obtained, and the content feature vector is input into the preset classification prediction model to perform content feature analysis to obtain the video feature information.

4. The video push method according to claim 1, characterized in that: The step of obtaining platform characteristic information and determining key characteristic information according to the video characteristic information and the platform characteristic information specifically includes: Acquire platform communication characteristic data, and perform feature extraction based on the platform communication characteristic data to obtain the platform characteristic information; Matching the video feature information with the platform feature information to obtain platform matching information; The video feature information is screened according to the platform matching information to obtain the key feature information.

5. The video push method according to claim 1, characterized in that: The step of reconstructing the initial video data based on the key feature information to generate valid video data specifically includes: Extract key information from the initial video data according to the key feature information to obtain key segments and important elements; Reconstruct the video based on the genetic algorithm, the key segments, and the important elements to generate a reconstructed video; Obtaining platform requirement information, and determining whether the video content of the reconstructed video meets the platform requirement information; If the video content of the reconstructed video meets the platform requirement information, the reconstructed video is used as the valid video data.

6. The video push method according to claim 5, characterized in that: The step of reconstructing the video based on the genetic algorithm, the key segments, and the important elements to generate the reconstructed video specifically includes: Acquire platform videos, and cluster them according to the playback performance of the platform videos to obtain performance video classes; Calculate the comprehensive performance score of the performance video category, and perform matching and screening in the performance video category according to the key segments, the important elements, and the comprehensive performance score to obtain the optimal platform video; The performance features of the platform video are extracted, and based on a genetic algorithm, the performance features of the platform video are used as an initial population, and the optimal platform video is used as a reference to perform video reconstruction to generate the reconstructed video.

7. The video push method according to claim 6, characterized in that: The step of obtaining platform videos and clustering the platform videos according to their playback performance to obtain performance video classes specifically includes: Get the playback performance data of platform videos; Extracting key elements from the platform video and establishing a video feature vector based on the key elements; The platform videos are divided into different performance video classes according to a clustering algorithm and the video feature vector.

8. A video push device, characterized in that: include: A data acquisition module, used for acquiring initial video data; A feature analysis module, used to perform feature analysis on the initial video data to obtain video content features; A feature analysis module, configured to perform content feature analysis on the initial video data based on the video content features to obtain video feature information; An attribute acquisition module, used to acquire platform characteristic information, and determine key characteristic information according to the video characteristic information and the platform characteristic information; The video push module is used to reconstruct the initial video data based on the key feature information, generate valid video data, and push the valid video data to the corresponding platform for video playback.

9. A computer device, characterized in that: The method comprises a memory and a processor, wherein the memory stores computer-readable instructions, and the processor implements the steps of the video pushing method according to any one of claims 1 to 7 when executing the computer-readable instructions.

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

Images