Video processing method, electronic equipment and storage medium

Abstract

The invention provides a video processing method, electronic equipment and a storage medium, and relates to the field of artificial intelligence technology and video processing. The method comprises the steps that a plurality of explanation videos of a first object are acquired, and different explanation videos are videos for explaining the first object in different explanation scenes; based on user behavior data of different explanation scenes, determining a target explanation video from the plurality of explanation videos, the user behavior data being used for representing behavior data generated by participating users of different explanation scenes; and associating the target explanation video to a description page of the first object, the description page being used for providing description information of the first object. The technical problem that the recommendation effect of the explanation video is poor in the related technology is solved.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of artificial intelligence and video processing, in particular to a video processing method, an electronic device and a storage medium. BACKGROUND

[0002] In the existing live e-commerce environment, the distribution mode of product explanation content is relatively single, usually relying on live recording videos or simply relying on the explanation of the last live broadcast. However, this simple relying method ignores the explanation quality and other factors, which may result in that the video explanation does not have enough attraction or conversion effect, thereby causing poor effect of product explanation video.

[0003] At present, no effective solution has been proposed for the above problems. SUMMARY

[0004] Embodiments of the present application provide a video processing method, an electronic device and a storage medium, to at least solve the technical problem of poor recommendation effect of explanation video in related technologies.

[0005] According to an aspect of an embodiment of the present application, a video processing method is provided, including: obtaining a plurality of explanation videos of a first object, wherein different explanation videos are videos for explaining the first object in different explanation scenes; determining a target explanation video from the plurality of explanation videos based on user behavior data of different explanation scenes, wherein the user behavior data is used to represent behavior data generated by participating users in different explanation scenes; and associating the target explanation video to a description page of the first object, wherein the description page is used to provide description information of the first object.

[0006] According to an aspect of an embodiment of the present application, a video processing method is provided, including: obtaining a plurality of explanation videos of a first object by calling a first interface, wherein different explanation videos are videos for explaining the first object in different explanation scenes, the first interface includes a first parameter, and a parameter value of the first parameter includes the plurality of explanation videos; determining a target explanation video from the plurality of explanation videos based on user behavior data of different explanation scenes, wherein the user behavior data is used to represent behavior data generated by participating users in different explanation scenes; associating the target explanation video to a description page of the first object, wherein the description page is used to provide description information of the first object; and outputting the target explanation video by calling a second interface, wherein the second interface includes a second parameter, and a parameter value of the second parameter includes the target explanation video.

[0007] According to another aspect of an embodiment of the present application, a computing device is also provided, including: a memory storing an executable program; and a processor configured to run the program, wherein the program performs the method in various embodiments of the present application when running.

[0008] According to a further aspect of the embodiments of the present application, an electronic device is provided, including: a memory storing an executable program; a processor connected with the memory through a bus, configured to run the program, wherein the program is configured to execute the method in the embodiments of the present application when running.

[0009] According to a further aspect of the embodiments of the present application, a computer readable storage medium is provided, including a stored executable program, wherein the computer readable storage medium is configured to control a device where the computer readable storage medium is located to execute the method in the embodiments of the present application when the executable program runs.

[0010] According to a further aspect of the embodiments of the present application, a computer program product is provided, including a computer program, which is configured to implement the method in the embodiments of the present application when executed by a processor.

[0011] According to a further aspect of the embodiments of the present application, a computer program product is provided, including a non-volatile computer readable storage medium storing a computer program, which is configured to implement the method in the embodiments of the present application when executed by a processor.

[0012] According to a further aspect of the embodiments of the present application, a computer program is provided, which is configured to implement the method in the embodiments of the present application when executed by a processor.

[0013] In the embodiments of the present application, a plurality of explanation videos of a first object are acquired, wherein different explanation videos are videos for explaining the first object in different explanation scenes; a target explanation video is determined from the plurality of explanation videos based on user behavior data of the different explanation scenes, wherein the user behavior data is used to represent behavior data generated by participating users in the different explanation scenes; and the target explanation video is associated to a description page of the first object, wherein the description page is used to provide description information of the first object. By analyzing the user behavior data, the attractiveness and conversion ability of each explanation video are quantitatively evaluated, and the better-performing explanation video is selected as the target explanation video after comparison, which is embedded into the description page of the first object to enhance the interactivity and recommendation effectiveness of the page, thereby solving the technical problem of poor recommendation effect of explanation videos in the related art.

[0014] The general description above and the detailed description below are intended to illustrate and explain the present application, and do not constitute a limitation on the present application. BRIEF DESCRIPTION OF DRAWINGS

[0015] The accompanying drawings, which are included to provide a further understanding of the application and are incorporated in and constitute a part of this application, illustrate embodiments of the application and together with the description serve to explain the application. In the drawings:

[0016] Figure 1 is a schematic diagram of an application scenario of a video processing method according to an embodiment of the application;

[0017] Figure 2 is a flowchart of a video processing method according to an embodiment of the application;

[0018] Figure 3 is a schematic diagram of a delivery process according to an embodiment of the application;

[0019] Figure 4 is a flowchart of a video processing method according to an embodiment of the application;

[0020] Figure 5 is a schematic diagram of a video processing device according to an embodiment of the application;

[0021] Figure 6 is a schematic diagram of a video processing device according to an embodiment of the application;

[0022] Figure 7 is a structural block diagram of a computing device according to an embodiment of the application;

[0023] Figure 8 is a structural block diagram of an electronic device according to an embodiment of the application. DETAILED DESCRIPTION

[0024] In order to enable persons skilled in the art to better understand the application scheme, the technical solutions in the embodiments of the application will be described clearly and completely below with reference to the drawings in the embodiments of the application. Obviously, the following described embodiments are only part of the embodiments of the application, rather than all the embodiments of the application. Based on the embodiments in the application, all other embodiments obtained by those skilled in the art without creative labor should belong to the scope of protection of the application.

[0025] The terms "first," "second," etc., used in the specification, claims, and accompanying drawings of this application are used to distinguish similar objects and are not necessarily used to describe a specific order or sequence. It should be understood that such data can be interchanged where appropriate so that the embodiments of this application described herein can be implemented in other orders. Wherein, "other orders" refers to orders other than those illustrated or described herein. Furthermore, the terms "comprising" and "having," and any variations thereof, are intended to cover a non-exclusive inclusion, such as a process, method, system, product, or apparatus that comprises a series of steps or units, not necessarily limited to those explicitly listed, but may include other steps or units not explicitly listed, or inherent to such processes, methods, products, or apparatus.

[0026] The user information (including but not limited to user device information, user personal information, etc.) and data (including but not limited to data used for analysis, stored data, displayed data, etc.) involved in this application are all information and data authorized by the user or fully authorized by all parties. Furthermore, the collection, use and processing of related data must comply with the relevant laws, regulations and standards of the relevant countries and regions, and corresponding operation portals are provided for users to choose to authorize or refuse.

[0027] The technical solution provided in this application is mainly implemented using a deep learning model. Deep learning models can be widely applied in fields such as Natural Language Processing (NLP), computer vision, and speech processing. Specifically, they can be applied to computer vision tasks such as Visual Question Answering (VQA), Image Captioning (IC), and image generation, as well as to natural language processing tasks such as text-based sentiment classification, text summarization, and machine translation. Therefore, the main application scenarios of this application include, but are not limited to, digital assistants, intelligent robots, search, online education, office software, e-commerce, and intelligent design.

[0028] First, some nouns or terms that appear in the description of the embodiments of this application shall be interpreted as follows:

[0029] Live product demonstration, or simply demonstration, refers to a live stream where the host introduces and showcases the functions and features of a particular product or product category to the audience through real-time video. This video segment is recorded and then displayed on the corresponding product page in the "Baby Pocket" or "Little Yellow Cart" app. The aim is to help viewers better understand the product, stimulate their interest in purchasing, and thus drive sales conversion.

[0030] "Affiliation" refers to a situation where a streamer has recorded a presentation about a product in a previous live stream, but no presentation was recorded for this live stream. However, this presentation can still be uploaded to the product page in this live stream.

[0031] An ab experiment involves changing only one variable while keeping all other conditions identical, and observing the impact of that variable on key metrics (such as click-through rate, conversion rate, and sales volume).

[0032] According to an embodiment of this application, a video processing method is provided. The steps shown in the flowcharts of the accompanying drawings can be executed in a computer system such as a set of computer-executable instructions, and although a logical order is shown in the flowcharts, in some cases the steps shown or described may be performed in a different order than that shown here.

[0033] The technical solutions provided in this application can employ deep learning models with relatively large parameter scales, such as large models containing billions or even more model parameters. Here, "large model" is just one example; this application does not limit the number of model parameters supported by the deep learning model used, aiming to meet actual needs. The deep learning models involved in this application can be artificial intelligence-based language models (LM) or multimodal models (MM).

[0034] Considering the limited computing resources of mobile terminals, the methods described above in this application embodiment can be applied to, for example... Figure 1 The application scenarios shown. Figure 1 This is a schematic diagram illustrating an application scenario of a video processing method according to an embodiment of this application. Figure 1 In the application scenario shown, the deep learning model is deployed on server 10. Server 10 can connect to one or more client devices 20 via a local area network (LAN), wide area network (WAN), internet connection, or other types of data network. Client devices 20 may include, but are not limited to, smartphones, tablets, laptops, PDAs, personal computers, smart home devices, and in-vehicle devices. Client devices 20 can interact with users through a graphical user interface to invoke the deep learning model, thereby implementing the method provided in this embodiment.

[0035] In this embodiment, the system consisting of a client device and a server can perform the following steps: The client device acquires multiple explanatory videos of a first object. The server, based on user behavior data from different explanatory scenarios, determines a target explanatory video from the multiple explanatory videos; and associates the target explanatory video with the description page of the first object.

[0036] With the rapid development of high-performance computing units, the methods provided in this application embodiment can also be applied to model-in-the-loop machines in other application scenarios. In one optional embodiment, the model-in-the-loop machine has multiple built-in models. Users can select a model to adjust as needed to obtain their own model. The high-performance computing unit built into the model-in-the-loop machine can then directly call the adjusted model to execute the methods provided in this application embodiment. In another optional embodiment, the deep learning model-in-the-loop machine has a pre-trained model built-in. The high-performance computing unit built into the model-in-the-loop machine can then directly call this model to execute the methods provided in this application embodiment.

[0037] Furthermore, when users need to train their own models, they can upload their own datasets via the client. This dataset is sent from the client to the server. The server can then use this dataset to fine-tune the pre-trained model, resulting in the user's customized model, which can then be deployed to the production environment. To facilitate user adjustments, the server provides complete adjustment tools, development frameworks, and processes, supporting various adjustment strategies. This allows the adjusted model to better adapt to different application domains and achieve a high degree of customization.

[0038] Under the aforementioned operating environment, this application provides the following: Figure 2 The video processing method shown. Figure 2 This is a flowchart of a video processing method according to an embodiment of this application. Figure 2 As shown, the method may include the following steps:

[0039] Step S202: Obtain multiple explanatory videos for the first object.

[0040] Among them, the different explanatory videos are videos that explain the first object in different explanatory scenarios.

[0041] The first object mentioned above can be the target product, which refers to a specific commodity or commodity category. This is the core content that the explanatory video focuses on. The first object can be a physical object, such as a commodity, or a virtual object, such as a digital collectible.

[0042] The aforementioned multiple explanatory videos refer to multiple video materials produced around the same primary object in different live-streaming scenarios, with each video reflecting a different live-streaming style or product selling point.

[0043] In one alternative embodiment, the technical actions involved in acquiring multiple explanatory videos of the first object include extracting all explanatory videos related to the first object from historical data of the live streaming platform. This typically involves automatic identification and segmentation techniques for the live streaming content to ensure that the acquired explanatory videos do indeed contain a detailed introduction to the first object. For example, the live streaming platform's algorithm can identify which video segments are about the first object through keyword matching, image recognition, or association with product inventory.

[0044] In the live-streaming e-commerce environment, hosts may introduce the same product multiple times across different live streams. Each live stream may generate multiple versions of the promotional video due to differences in live stream duration, audience composition, and host's style. This application can automatically or semi-automatically identify and extract all relevant promotional videos from these historical live stream data, serving as the basis for subsequent analysis and selection. This process may include technologies such as data mining, video analysis, and intelligent recognition.

[0045] This application systematically collects all explanatory video resources related to the first object, providing a comprehensive data foundation for subsequent analysis and selection based on user behavior data. Through automatic identification and slicing technology, it greatly improves data collection efficiency while avoiding the subjectivity and omissions of manual screening, ensuring the accuracy and comprehensiveness of subsequent analysis, and effectively improving the utilization efficiency of explanatory videos and the operational quality of live-streaming e-commerce.

[0046] Step S204: Based on user behavior data from different explanation scenarios, determine the target explanation video from multiple explanation videos.

[0047] Among them, user behavior data is used to represent the behavioral data generated by participating users in different explanation scenarios.

[0048] The aforementioned user behavior data covers users' reactions to explanatory videos and their interactions with products in specific explanatory scenarios, such as clicks, dwell time, and purchase behavior.

[0049] The aforementioned selection of target instructional videos is based on the results of user behavior data analysis, choosing those instructional videos that demonstrate high click-through rates, sales volume, and conversion rates within a unit of time.

[0050] This application enables the rapid and accurate identification and extraction of explanatory videos related to the primary subject from massive amounts of live streaming data. These explanatory videos cover different live stream sessions and broadcaster styles, providing diverse explanatory perspectives. Through automated video analysis and segmentation, not only are labor costs saved, but data integrity is also ensured.

[0051] The core of this application focuses on the selection and improvement of explanatory videos. Its main objective is to identify target explanatory videos—those that perform well in different explanatory scenarios—based on user behavior data. This process is a crucial step in data-driven evaluation of the effectiveness of live-streamed explanation videos, ensuring the user appeal of the video content.

[0052] First, the system collects behavioral data from participating users in each presentation scenario. This data includes, but is not limited to, viewer clicks, purchases, viewing time, interactions (such as comments, likes, and shares), and subsequent return rates and user feedback. Data collection needs to span multiple dimensions to ensure a comprehensive reflection of user reactions to the presentation videos. Next, the system evaluates the collected user behavioral data, quantifying these behaviors into numerical metrics.

[0053] For example, click-through rate per unit time reflects the video's ability to attract viewers' attention, while sales volume per unit time reflects the video's sales conversion efficiency. Conversion rate measures the persuasiveness of the video, completion rate and interaction rate (likes, comments, shares) help judge the content's appeal and user engagement, while return rate and after-sales feedback are used to verify the credibility of the content. The system maps the quantified user behavior data to video ratings based on a pre-set scoring model and a weighted comprehensive evaluation strategy. The scoring model can employ the analytic hierarchy process (AHP) or machine learning ranking models to ensure objectivity and accuracy. The scoring model is improved by combining A / B testing results, scoring each video and calculating the total rating for all videos. Finally, the system compares the ratings of all videos and identifies the videos with higher ratings as target videos.

[0054] By collecting and evaluating multi-dimensional behavioral data from participating users, and using a pre-defined scoring model to quantify these behaviors into scores, high-performing instructional videos are identified. This process not only quantifies the market performance and user engagement of instructional videos but also ensures that high-quality content is accurately identified for inclusion on product pages and for intelligent recommendations.

[0055] Step S206: Associate the target explanation video with the description page of the first object.

[0056] The description page is used to provide descriptive information about the first object.

[0057] The description page of the first object mentioned above can be a product description page. The product description page provides an interface with detailed information about the product, including but not limited to images, text descriptions, specifications, etc., and is one of the main ways for users to learn about the product.

[0058] The aforementioned association can refer to the process of embedding or linking the target explanation video to the product description page, so that users can directly access the relevant explanation video when viewing product details, thereby enhancing the intuitiveness and attractiveness of product information.

[0059] The aforementioned target instructional videos refer to instructional videos that have been identified as performing well after analyzing user behavior data, and are used for display on subsequent product description pages.

[0060] The aforementioned description page is a webpage that presents users with detailed product information, including but not limited to product images, text descriptions, specifications, etc., and is an important reference in the purchase decision-making process.

[0061] This application can link selected instructional videos to product description pages through linking or embedding technology, ensuring that users can directly access the video when browsing product information.

[0062] The key step in this application embodiment, which links the selected explanatory video with the product description page, aims to enhance the attractiveness and persuasiveness of product information through intuitive video content, thereby improving the efficiency of users' purchase decisions. This process is an important way to improve content and enhance user experience in live-streaming e-commerce.

[0063] The system has identified target instructional videos based on user behavior data—videos with high market performance and user engagement among multiple instructional videos. The system then locates the product description page, the primary display area for product information, typically containing product images, detailed descriptions, prices, and a purchase button, aiming to provide comprehensive product information to potential buyers. Next, the system embeds the target instructional video into the product description page, usually as part of the "Recommended Video" or "Product Explanation" section. This embedding can be achieved by embedding a video player, adding a video link, or directly uploading the video to the page. The system may also need to adjust the layout of the product description page to ensure the target instructional video is prominently displayed without affecting page loading speed or user browsing experience. The video's display position, size, and playback settings need to be confirmed to enhance user attention and interaction. By linking the target instructional video to the product description page, the system achieves content reuse and multi-channel distribution, not only enriching the product information.

[0064] For example, linking target explanation videos with product description pages is an effective way to improve product conversion rates and user shopping experience. The system locates the primary product description page and embeds high-quality, market-proven explanation content into it, creating an intuitive and interactive product display method. For instance, after host J provides a vivid explanation of a smartwatch, the system identifies this explanation video as the target explanation video based on user click-through rate, conversion rate, and other data. Subsequently, the system embeds this video into the smartwatch's product details page as part of the "Host Recommendation" section, significantly enhancing the page's attractiveness and user engagement. Simultaneously, the system can adjust the page layout and video display strategy based on the video's actual performance to ensure a better user experience. This process not only improves the way product information is disseminated but also achieves the goal of "one live stream, multiple monetizations" through content reuse, significantly improving the operational efficiency and profitability of live-streaming e-commerce.

[0065] For example, in the live-streaming e-commerce sector, users may miss detailed explanations of certain products while watching a live stream, or prefer to browse product details in a silent state. Linking the target product demonstration video to the description page means that in the subsequent purchase process, even if users haven't directly watched the live stream, they can directly access the best-performing demonstration video on the product details page. This not only enriches the way product information is displayed but also provides more information for purchasing decisions. For instance, when users search for or browse products, they can instantly see live-stream demonstration videos related to the product, increasing the product's intuitive appeal and trustworthiness.

[0066] Linking product demonstration videos with the initial product description page significantly enhances the richness and appeal of product information. This technology not only increases the interactivity of product pages but also strengthens users' understanding and confidence in the products through intuitive video explanations, helping to improve click-through rates and conversion rates, thereby driving sales growth. Simultaneously, it provides users with a more multi-dimensional shopping experience, compensating for the limitations of relying solely on text and image descriptions, promoting diversified communication of product information, and improving the operational effectiveness and user experience of live-streaming e-commerce.

[0067] Once the system identifies the target product demonstration video through user behavior data analysis, it can then effectively link that video to the product description page. This linking process goes beyond simply embedding a video link; it requires ensuring deep integration of the video and product information, as well as smooth video playback within the page. By embedding the target demonstration video into the description page, users gain more intuitive and vivid product information. This not only enriches the presentation of product information but also increases user interest and understanding. Users can watch the live demonstration directly on the product page without needing to navigate to a replay or other channels. This seamless experience helps improve user dwell time, click-through rates, and purchase conversion rates. It also improves the reuse of live stream content. Even after the live stream ends, high-quality demonstration videos can continue to drive traffic and conversions for the product, achieving a long-tail effect for live stream content and enhancing the operational efficiency and content value of live stream e-commerce.

[0068] This application aims to improve the efficiency of selecting and distributing product demonstration videos in live-streaming e-commerce, thereby promoting product sales and improving user experience. The system first collects all demonstration videos related to the first product from live-stream replays, recorded segments by the host, or other sources. These videos come from different demonstration scenarios—different live-stream sessions, different hosts, and different audience sizes—ensuring the diversity of video content and scenario coverage. Next, the system analyzes user interaction behavior in different demonstration scenarios, including but not limited to viewer click-through rates, purchase conversion rates, viewing time, comments, and sharing data. Through quantitative evaluation, the system identifies demonstration videos that perform well in terms of user engagement and sales conversion as target demonstration videos. This process utilizes insights from user behavior data to ensure that the target demonstration videos accurately reflect user preferences and product appeal. Finally, the system embeds the selected target demonstration videos into the description page of the first product as intuitive display material for product information. This not only enriches the product description and increases the interactivity and appeal of the page but also helps users better understand the product's characteristics and uses through the intuitive demonstration of the live-streaming video, thereby increasing purchase intention and conversion rates.

[0069] Through this series of operations, the system not only effectively selected instructional videos with positive market feedback but also enhanced the intuitiveness and appeal of product information by linking them to product description pages. This improved the efficiency of user purchase decisions, thereby driving sales conversion for smartwatches and improving the operational effectiveness and user experience of live-streaming e-commerce. This case demonstrates the significant effect of this application's embodiments in improving the utilization efficiency of product instructional videos and the operational effectiveness of live-streaming e-commerce.

[0070] Through the above steps, multiple explanatory videos for the first object are obtained. These videos explain the first object in different scenarios. Based on user behavior data from different scenarios, a target explanatory video is selected from these videos. This user behavior data represents the behavioral data generated by participating users in different scenarios. The target explanatory video is then associated with the description page of the first object, which provides descriptive information about the first object. By analyzing user behavior data, the attractiveness and conversion rate of each video are quantitatively evaluated. After comparison, the best-performing video is selected as the target video and embedded into the description page of the first object to enhance page interactivity and recommendation effectiveness. This solves the technical problem of poor recommendation performance of explanatory videos in related technologies.

[0071] In this embodiment of the application, determining the target explanation video from multiple explanation videos based on user behavior data of different explanation scenarios includes: using a scoring model to score multiple explanation videos based on user behavior data of different explanation scenarios to obtain score values ​​for multiple explanation videos, wherein the score values ​​are used to quantify the explanation quality of the first object in different explanation videos; and filtering multiple explanation videos based on the score values ​​of multiple explanation videos to obtain the target explanation video.

[0072] The scoring model described above is used to quantitatively evaluate the performance of explanatory videos in different explanatory scenarios. By setting rules and weights, it transforms various user behavior data into comparable score values.

[0073] The aforementioned user behavior data records a series of behavioral indicators of users when watching instructional videos and browsing product pages, such as click-through rate, viewing time, purchase rate, and interaction frequency. These data reflect users' interest in the video content and their acceptance of the products.

[0074] The above rating values ​​are the output of the rating model and are used to measure the quality and appeal of the explanatory video. The higher the value, the better the video performs in the corresponding explanatory scenario.

[0075] The above filtering process involves selecting items that meet certain criteria or conditions from multiple options. In this case, it means ranking the videos based on their ratings and selecting the top-ranked videos as the target videos.

[0076] First, a rating model is built based on existing user behavior data. This model converts different user behavior metrics (such as click-through rate and conversion rate) into rating values, reflecting the performance of the instructional videos in different scenarios. Then, the rating model is applied to each instructional video to obtain its individual rating. This process can utilize techniques such as data cleaning, feature engineering, and algorithm optimization to ensure the accuracy and effectiveness of the ratings.

[0077] The process of filtering multiple instructional videos based on ratings to obtain target instructional videos is a rating-based sorting process. The system sorts the collected ratings and selects instructional videos with higher ratings. These videos are considered to perform better in explaining the primary subject matter, and are more effective in attracting users and improving conversion rates. Therefore, they are considered target instructional videos for subsequent affiliation and distribution.

[0078] In live-stream e-commerce, user reactions to live-stream explanation videos vary widely. Some videos attract numerous clicks due to their engaging content, while others achieve high conversion rates thanks to the host's presentation style. To objectively evaluate and select explanation videos that effectively improve product exposure and purchase rates, this application employs a scoring model for quantitative analysis. The scoring model comprehensively considers multiple dimensions of user behavior data, such as click-through rate, purchase rate, and viewing time, assigning a score to each explanation video. This allows for the selection of target explanation videos from numerous candidate videos, which can then be used for integration into product pages and multi-channel distribution.

[0079] This application achieves objective quantification and intelligent screening of explanatory video quality. Utilizing a scoring model, the system transforms abstract user behavior into concrete scoring indicators, providing data support for the performance of explanatory videos. The scoring-based screening mechanism ensures efficient selection of target explanatory videos, avoids the subjectivity of manual judgment, improves the accuracy and effectiveness of explanatory videos associated with product pages, and further promotes conversion rates and user satisfaction in live-streaming e-commerce.

[0080] In this embodiment, a scoring model is used to scientifically quantify and intelligently filter instructional videos. The scoring model fully considers key dimensions of user behavior data, assigning an objective score to each instructional video by comprehensively evaluating metrics such as click-through rate, conversion rate, and viewing time. This accurately reflects the video's appeal and sales potential. Based on the score-based filtering, the system automatically selects the best-performing instructional videos as target videos for inclusion on product pages and for subsequent live stream recommendations. This technical action significantly improves the efficiency of product information display and the operational efficiency of live-streaming e-commerce. Users can watch selected high-quality instructional videos on product description pages, increasing their understanding and interest in the products and improving the efficiency of their shopping decisions. This significantly increases click-through rates and purchase conversion rates, bringing user satisfaction to live-streaming e-commerce.

[0081] In this embodiment of the application, a scoring model is used to score multiple explanation videos based on user behavior data in different explanation scenarios, thereby obtaining multiple score values ​​for the explanation videos. This includes: extracting features from user behavior data in multiple feature extraction dimensions to obtain multiple user behavior features; determining target features for different explanation scenarios based on the feature weights corresponding to the multiple feature extraction dimensions and the multiple user behavior features; and determining the score values ​​for the multiple explanation videos based on the target features for different explanation scenarios.

[0082] The aforementioned feature extraction dimensions are multiple perspectives or indicators set in data analysis to quantify and evaluate user behavior data, such as click-through rate, viewing time, conversion rate, and interaction rate.

[0083] The aforementioned user behavior characteristics are extracted by analyzing user behavior data in scenarios such as live streaming, and from this data, feature data reflecting user preferences, interests, and purchase intentions are obtained.

[0084] The aforementioned feature weights are a quantitative expression of the degree of influence of different feature extraction dimensions on the final score in the scoring model, used to adjust the relative importance of each dimension in the evaluation process.

[0085] The aforementioned target features are key features used to evaluate the quality of narration videos, determined based on feature weights and user behavior characteristics, under different narration scenarios.

[0086] This application can extract meaningful features from raw user behavior data. For example, the system can analyze the time users spend in the live stream, the number of times they click on product links, and the frequency of product purchases, converting this data into features such as click-through rate and viewing duration. Feature weights are set to highlight the importance of different metrics in evaluating instructional videos. The system can dynamically adjust feature weights based on the desired objectives and historical data analysis, ultimately determining representative target features for a specific instructional scenario. After obtaining the target features, the system uses a scoring model to score each instructional video, matching the video's performance with the target features to derive a score. The scoring model can include various methods such as weighted averaging and machine learning algorithms to ensure the comprehensiveness and accuracy of the scoring.

[0087] In the live-streaming e-commerce sector, platforms need to sift through massive amounts of live-streaming data to identify instructional videos that can effectively boost product sales and user engagement. First, by extracting features from user behavior data across multiple dimensions, the system can identify user behavior patterns such as clicks, views, and purchases during live streams. Next, feature weights are assigned based on the desired objectives to determine which dimensions of user behavior are most critical for evaluating the quality of the instructional videos. Finally, based on these target features, a scoring model assigns scores to the instructional videos. High-scoring videos are considered target instructional videos that significantly promote product sales and user interaction and are prioritized for recommendation and distribution.

[0088] By implementing the feature extraction and scoring model described in this application, live streaming platforms can more accurately identify which explanatory videos have high appeal and conversion potential based on data-driven decision-making. This technical process not only improves the efficiency and accuracy of explanatory video selection but also promotes the effective dissemination of product information, thereby improving the user experience and sales conversion rate of live streaming e-commerce.

[0089] In this embodiment, by conducting in-depth analysis of user behavior data across multiple feature extraction dimensions, the system can capture users' genuine interest in and purchase intentions regarding the instructional videos, thereby extracting multiple user behavior features. Based on these features, and combined with the feature weights set for the desired objectives, target features for different instructional scenarios are determined, providing clear indicators and weights for the construction of the scoring model. Ultimately, the scoring model achieves a quantitative evaluation of the instructional video quality by rating the videos, helping to accurately identify target instructional videos from numerous candidate videos that can significantly increase product sales and user interaction. The implementation of this technical process improves the utilization efficiency of live streaming content, enhances the way product information is delivered, and promotes higher sales conversion rates in live streaming e-commerce.

[0090] In this embodiment of the application, the rating values ​​of multiple explanation videos are determined based on the target features of different explanation scenarios, including: mapping the target features of different explanation scenarios based on a preset correspondence to obtain the rating values ​​of multiple explanation videos, wherein the preset correspondence is used to represent the correspondence between different features and different rating values.

[0091] The aforementioned pre-defined correspondence is a set of rules or mapping tables used to connect features extracted from user behavior data with their corresponding rating values. This relationship ensures the accurate reflection of different features in the ratings, as well as the consistency and comparability of the ratings.

[0092] For example, user behavior characteristics of multiple explanatory videos are identified, and these characteristics are mapped to predefined rating values ​​based on a preset correspondence. For instance, a 0.01% increase in click-through rate per unit time results in a 1-point increase in rating; a 100 yuan increase in transaction volume per unit time results in a 2-point increase in rating; and a 0.1% increase in conversion rate results in a 5-point increase in rating. This mapping process ensures that the conversion of feature data and the generation of rating values ​​follow established rules. After feature mapping is completed, the target features of each explanatory video (click-through rate per unit time, transaction volume per unit time, conversion rate, etc.) are converted into a set of rating values ​​through the preset correspondence. These rating values ​​comprehensively reflect the performance of the explanatory video in different explanatory scenarios.

[0093] For example, in the context of live-streaming e-commerce, the quality of the instructional videos directly affects the product's exposure, conversion rate, and user engagement. This application embodiment, by setting a preset correspondence, transforms user behavior data while watching instructional videos into intuitive rating values. This allows the system to automatically identify which videos are more popular and have better conversion rates, thereby providing data support for product page integration and intelligent recommendations in subsequent live streams.

[0094] By implementing the feature mapping and scoring generation steps of this application embodiment, the live streaming platform can effectively quantify the performance of different explanatory videos and achieve an objective evaluation of the quality of the explanatory videos. This not only improves the efficiency of explanatory video selection but also ensures that the distributed videos can more effectively promote product conversion and improve user satisfaction, thereby enhancing the operational efficiency and content value of live streaming e-commerce.

[0095] In this embodiment, the pre-defined correspondence ensures that user behavior data can be accurately converted into rating values, reflecting the performance of the instructional videos in different instructional scenarios. For example, target features such as click-through rate per unit time, transaction volume per unit time, conversion rate, and completion rate are assigned corresponding rating values ​​through the pre-defined correspondence. These rating values ​​comprehensively reflect the attractiveness and conversion efficiency of the videos. Based on the rating values, the system can automatically identify and select instructional videos with excellent market performance as target instructional videos, linking them to product description pages and providing intelligent recommendations for subsequent live streams. This technical process not only improves the accuracy and efficiency of content selection but also promotes the effective dissemination of product information, enhances user engagement, and ultimately drives sales conversion and brand influence in live e-commerce, improving the reusability and operational efficiency of live content.

[0096] In this embodiment of the application, multiple explanatory videos are filtered based on their ratings to obtain a target explanatory video, including: determining the explanatory video with the highest rating among the multiple explanatory videos as the target explanatory video based on their ratings.

[0097] The rating values ​​for the aforementioned instructional videos are numerical values ​​obtained by quantitatively evaluating the performance of the instructional videos through a rating model, reflecting key indicators such as the attractiveness and conversion rate of the instructional videos in different instructional scenarios.

[0098] The highest score mentioned above represents the highest score among a group of explanatory videos, indicating that the video with the highest score in that group is the best performing one.

[0099] The aforementioned target instructional videos were selected based on their ratings and are considered high-performing. These videos will be used for subsequent integration into product pages and intelligent recommendations.

[0100] All instructional videos are sorted by rating from highest to lowest to facilitate identification of the video with the highest rating. From the sorted list of video ratings, the video with the highest rating is selected, indicating it performed well among all candidate videos. Based on this highest rating, the video is designated as the target instructional video, signifying its strong performance within the current evaluation period and its suitability for integration into product pages and intelligent recommendations during subsequent live streams.

[0101] In the operation of live-streaming e-commerce, platforms need to regularly evaluate and select high-performing instructional videos to improve product exposure and conversion rates. This application's embodiment achieves objective comparison between instructional videos through quantified scoring, thereby accurately identifying and selecting target instructional videos during product page placement and live-streaming recommendation, effectively enhancing the reuse value and conversion efficiency of live-streaming content.

[0102] This application can automatically identify and determine high-performing instructional videos based on quantitative scoring, significantly improving the utilization efficiency and conversion rate of live-streaming content. This technology simplifies the content selection process in live-streaming e-commerce, avoids the subjectivity and inconsistency of manual judgment, ensures a more objective and efficient selection of target instructional videos, and thus improves the way product information is delivered.

[0103] For example, live-streaming e-commerce platforms can automatically identify and determine high-performing target videos based on their ratings. This technology not only simplifies the content selection process and improves efficiency but also ensures objective evaluation of video quality, avoiding the subjective biases of traditional manual selection. For instance, when evaluating a newly launched smart headphone, the system uses a rating model to score the demonstration videos of the same product across multiple live streams, identifying videos with higher ratings and using them as target videos for intelligent recommendation. This process not only enhances the attractiveness and conversion rate of product pages but also strengthens the interactivity and user engagement of the live stream through the intelligent recommendation system, leading to higher sales and user satisfaction for live-streaming e-commerce. This demonstrates the crucial role of this solution in improving the reusability and operational efficiency of live-streaming content.

[0104] In this embodiment of the application, the method further includes: outputting multiple explanatory videos and rating values ​​for the multiple explanatory videos; and updating the model parameters of the rating model based on the feedback information received regarding the rating values ​​of the multiple explanatory videos, wherein the feedback information is obtained by adjusting the rating values ​​of the multiple explanatory videos according to the multiple explanatory videos.

[0105] The feedback information mentioned above includes user feedback on the instructional videos, actual performance data of the videos (such as actual conversion rate, return rate, etc.), and expert suggestions for adjusting the scoring model. These are used to evaluate the accuracy and effectiveness of the scoring model and guide the updating of the scoring model parameters.

[0106] Multiple instructional videos and their corresponding ratings can be presented to users or administrators for easy viewing and analysis. Feedback on the actual market performance of these videos can be collected, including but not limited to return rates, actual conversion rates, and product reviews, to verify the accuracy of the rating model's predictions. Based on this feedback, the parameters of the rating model can be adjusted to better reflect real user feedback and market performance, thereby improving rating accuracy and model performance.

[0107] For example, in a live-streaming e-commerce environment, the system's rating model may need continuous improvement and adjustment to adapt to market changes and user preferences. First, the system outputs the rating score for the instructional video, allowing the operations team or the streamer to review the evaluation results. Next, the system receives follow-up feedback on the instructional video and analyzes how well this feedback matches the previous rating score. Finally, based on the feedback, the parameters of the rating model are updated, making the rating mechanism more aligned with users' actual needs and market trends.

[0108] The operation of this application embodiment not only makes the video scoring process transparent, improving the fairness and transparency of live e-commerce content evaluation, but also allows for dynamic updates to the scoring model based on real feedback information, improving the model's accuracy and reliability, thereby promoting the improvement of live content and the overall operational efficiency of live e-commerce.

[0109] In this embodiment, the explanatory video and rating are output to the operator, such as the live streaming platform's operations team or the streamer, allowing them to intuitively understand the video's performance and promoting self-review and improvement of the content. Receiving feedback and updating key parameters of the rating model accordingly ensures the rating system can adapt to market changes and user preferences in a timely manner, avoiding rating lag. For example, if an explanatory video receives a high rating due to its high content quality but fails to convert into sales in actual application, or even leads to an increased return rate, the feedback will highlight this mismatch, prompting the system to correct the parameters of the rating model to more accurately reflect the video's market value. By implementing the above operations, this embodiment achieves self-improvement of the rating mechanism, improves the accuracy of the rating and the adaptability of the model, thereby enhancing the efficiency of live streaming content selection and the operational effectiveness of live streaming e-commerce.

[0110] In this embodiment of the application, obtaining multiple explanatory videos of the first object includes: obtaining multiple target videos, wherein different target videos are videos in which the second object gives explanations in different explanation scenarios; performing content detection on the multiple target videos to obtain the explanation time period of the first object; and slicing the multiple target explanatory videos according to the explanation time period to obtain multiple explanatory videos.

[0111] The aforementioned target videos are video clips in live-streaming e-commerce where the host explains specific products or services. These videos can be included in the live-stream replay or are excerpts of explanations specially recorded by the host during the live stream.

[0112] The aforementioned explanation time period refers to the time segment in the target video during which the host provides a detailed description and presentation of the first object. Content detection technology can automatically or manually identify this time segment for subsequent video slicing processing.

[0113] The system retrieves all target videos where the host explains different products from live stream replays or other content libraries, providing raw data for subsequent video slicing and analysis. By applying audio and video analysis technologies, it identifies the explanation content related to the primary product within the target videos and determines the explanation time periods, used to accurately locate the start and end points of video segments. Based on the identified explanation time periods, the target videos are segmented into multiple explanation videos, each containing complete explanation content for the primary product, facilitating subsequent quantitative evaluation and affiliate distribution.

[0114] For example, in the operation of live-streaming e-commerce, the system needs to efficiently extract video clips explaining a specific product from a massive amount of live-stream replays. First, the system acquires multiple target videos, covering the host's explanations of similar products at different times and in different scenarios. Next, the system uses content detection technology to identify the specific time period for explaining the primary product. Finally, based on the identified explanation time period, the target video is segmented into multiple explanation videos, each focusing on the explanation of one product, facilitating subsequent quantitative analysis and content distribution.

[0115] This application can accurately extract video clips containing product explanations from live stream replays, providing precise raw data for subsequent quantitative analysis and content distribution, thereby improving the utilization efficiency and distribution accuracy of live stream content.

[0116] In this embodiment of the application, the method further includes: if the first object has a new explanation scenario, during the explanation process of the second object for the first object, outputting the target explanation video in the client of the second object.

[0117] The second object mentioned above can be the subject explaining the first object, and the second object can be the live streamer in live e-commerce. The explanation scenario mentioned above refers to the specific environment or background in which the live streamer explains the product in live e-commerce. It can be different live stream sessions, live stream themes, or user groups. The explanation video in each scenario can vary depending on the live streamer's style, product characteristics, and audience reactions.

[0118] The aforementioned target explanation videos were identified based on system analysis as high-quality videos with good market performance. They are used as supplementary or reference videos in new explanation scenarios to improve explanation effectiveness and product conversion rates.

[0119] The aforementioned client, in the above scenario, refers to the device on which the user watches the live stream or product details page, such as a smartphone, computer, or tablet, which is the terminal through which the system outputs content to the user.

[0120] When the system detects the first object appearing in a new explanation scenario, it pushes this video clip to the client of the broadcaster who is explaining the product, based on the previously analyzed target explanation video. This push process can serve as reference material for the broadcaster before the live stream begins, or as a real-time auxiliary tool during the live stream, helping the broadcaster improve the quality and effectiveness of their explanation. Furthermore, the target explanation video can also be directly embedded in the product details page for users to watch while browsing the product, increasing its appeal and encouraging purchase decisions.

[0121] In live-stream e-commerce operations, hosts often need to explain the same product, but the explanation style, target audience, and market environment can vary each time. This application's embodiment provides a valuable reference for hosts in new explanation scenarios by outputting a target explanation video to the client, enabling them to quickly understand which explanation methods and sales pitches are more popular, thereby improving live-stream performance and product conversion rates. Simultaneously, users can directly watch the target explanation video on the product details page, enhancing the intuitiveness and appeal of product information and improving the efficiency of user purchase decisions.

[0122] By implementing this step, the target explanation video can be accurately located and output to the host in the new explanation scenario or embedded in the product details page, which improves the content reuse efficiency and product conversion rate in live e-commerce. At the same time, it also improves the quality of the host's explanation and the user's shopping experience.

[0123] This application proposes a method for the optimal selection and intelligent distribution of live-stream product explanations. By conducting in-depth analysis of product explanation videos from the anchor's past live streams, and utilizing indicators such as click-through rate and transaction volume per unit time, it accurately identifies and selects explanations with attractiveness and sales conversion capabilities, enabling efficient attachment and distribution in subsequent live streams or product detail pages, thereby improving the reuse value and sales efficiency of live-stream content.

[0124] Quantitative evaluation ensures that every presentation is of verified quality, enhancing the professionalism and appeal of product descriptions. High-quality content is strategically placed on product pages or in live streams, significantly increasing viewer purchase intent and directly driving sales conversion, achieving "one live stream, multiple monetizations." Analysis of past presentations provides valuable insights for broadcasters, guiding them to employ more effective presentation techniques and scripts, continuously improving the quality and impact of their live streams. Avoiding repetitive recordings reduces production costs, while building a content asset library facilitates management and reuse, improving the operational efficiency of live-stream e-commerce. Presenting users with carefully selected content increases their trust and interest in the products, creating a better shopping experience.

[0125] Systematically analyzing recorded product presentation segments from different live streams by broadcasters, and selecting high-quality content based on key metrics such as Click-Through Rate (CTR), Gross Merchandise Volume (GMV), and Conversion Rate (CVR), is a crucial strategy for improving the operational efficiency and content reuse value of live-streaming e-commerce. This allows for efficient integration and multi-channel distribution. Presentations can be recorded by the broadcaster themselves for each live stream or generated by algorithms.

[0126] For each presentation segment, the scoring model was improved by combining A / B testing. Based on live stream backend and user behavior data, the following Key Performance Indicators (KPIs) were calculated to identify the best presentations. Click-through rate per minute measures the activity level of viewers clicking on product links or shopping carts during the segment, reflecting content appeal. Gross merchandise volume per minute (GMV per minute) assesses the direct sales contribution of the segment, reflecting conversion efficiency. Conversion rate determines the ratio of users placing orders to viewers, judging the persuasiveness of the presentation. Completion rate and interaction rate include likes, comments, and shares, helping to judge content appeal and user engagement. Return rate and after-sales feedback (optional) are used to verify the authenticity of the presentation and whether there are any over-promises.

[0127] A weighted comprehensive scoring model is used to compare all segments horizontally and select "high-value explanatory segments" that excel in multiple dimensions. The weighted comprehensive scoring model can be the Analytic Hierarchy Process (AHP) or a machine learning ranking model.

[0128] By attaching high-quality explanatory segments to product pages, these segments can be embedded into the product details page, short video showcase, or search results page as "host-recommended videos," improving user decision-making efficiency and trust. Artificial Intelligence (AI) intelligent recommendations allow the system to automatically recommend historically high-conversion scripts or segments for reference during subsequent live streams, assisting hosts in improving their real-time explanations. In this continuous iteration and improvement loop, a "recording-analysis-optimization-distribution-feedback" closed-loop mechanism can be established to continuously monitor the secondary conversion effect of distributed segments; guiding hosts to reuse high-performing script structures in subsequent live streams, thereby improving the overall quality of the live stream.

[0129] By employing the above strategies, we can not only maximize the content value of a single live stream, but also achieve "one live stream, multiple monetizations," significantly improving the efficiency of matching people, goods, and venues, and building a sustainable live stream content asset library for brands and platforms.

[0130] Figure 3 This is a schematic diagram illustrating the delivery process according to an embodiment of this application, such as... Figure 3 As shown, first, the host records a new explanation, and then conducts an A / B experiment with this explanation and the host's previous best explanations to compare the benefits of the two per unit time, obtain the key highlights, and then release the best explanation of this round. After recording a new explanation, the above steps are repeated continuously.

[0131] Through continuous A / B testing, this application enables the system to intelligently compare the historical and current video effects of the same host and product combination, automatically select and distribute better content, significantly improve the quality of the presentation and the sales conversion rate, and build an efficient closed loop for live streaming content distribution.

[0132] For the sake of simplicity, the foregoing method embodiments are all described as a series of actions. However, those skilled in the art should understand that this application is not limited to the described order of actions. This is because, according to this application, certain steps can be performed in other orders or simultaneously. Furthermore, those skilled in the art should also understand that the embodiments described in the specification are preferred embodiments, and the actions and modules involved are not necessarily essential to this application.

[0133] Through the above description of the embodiments, those skilled in the art can clearly understand that the methods according to the above embodiments can be implemented by means of software plus necessary general-purpose hardware platforms, or by hardware. Based on this understanding, the technical solution of this application, in essence, or the part that contributes to the prior art, can be embodied in the form of a software product. This computer software product is stored in a storage medium (such as ROM / RAM, magnetic disk, optical disk), and includes several instructions to cause a terminal device (which may be a mobile phone, computer, server, or network device, etc.) to execute the methods of the various embodiments of this application.

[0134] Figure 4 This is a flowchart of a video processing method according to an embodiment of this application, such as... Figure 4 As shown, the method may include the following steps:

[0135] Step S402: Obtain multiple explanatory videos of the first object by calling the first interface.

[0136] Among them, different explanatory videos are videos that explain the first object in different explanatory scenarios. The first interface includes a first parameter, and the parameter value of the first parameter includes multiple explanatory videos.

[0137] The aforementioned first interface can be an interface for data interaction between the cloud server and the client. Multiple explanatory videos can be passed into the interface function as the first parameter of the interface function to achieve the purpose of uploading multiple explanatory videos to the cloud server.

[0138] Step S404: Based on user behavior data from different explanation scenarios, determine the target explanation video from multiple explanation videos.

[0139] Among them, user behavior data is used to represent the behavioral data generated by participating users in different explanation scenarios.

[0140] Step S406: Associate the target explanation video with the description page of the first object.

[0141] The description page is used to provide descriptive information about the first object.

[0142] Step S408: Output the target explanation video by calling the second interface.

[0143] The second interface includes a second parameter, the value of which includes the target explanation video.

[0144] The aforementioned second interface can be an interface for data interaction between the cloud server and the client. The cloud server can pass the target explanation video into the interface function as the second parameter of the interface function, thereby achieving the purpose of sending the target explanation video to the client.

[0145] Through the above steps, multiple explanatory videos for a first object are obtained by calling a first interface. These videos explain the first object in different scenarios. The first interface includes a first parameter, the value of which includes multiple explanatory videos. Based on user behavior data from different scenarios, a target explanatory video is determined from these videos. This user behavior data represents the behavioral data generated by participating users in different scenarios. The target explanatory video is then associated with the description page of the first object, which provides descriptive information about the first object. Finally, the target explanatory video is output by calling a second interface, which includes a second parameter, the value of which includes the target explanatory video. By analyzing user behavior data, the attractiveness and conversion ability of each explanatory video are quantitatively evaluated. After comparison, the best-performing video is selected as the target explanatory video, which is then embedded into the description page of the first object to enhance page interactivity and recommendation effectiveness, thereby solving the technical problem of poor recommendation performance of explanatory videos in related technologies.

[0146] According to an embodiment of this application, a schematic diagram of a video processing apparatus for implementing the above-described video processing method is also provided. Figure 5 This is a schematic diagram of a video processing apparatus according to an embodiment of this application, such as... Figure 5 As shown, the device 500 includes: an acquisition module 502, a determination module 504, and an association module 506.

[0147] The acquisition module is used to acquire multiple explanatory videos of the first object, wherein different explanatory videos are videos that explain the first object in different explanatory scenarios; the determination module is used to determine the target explanatory video from the multiple explanatory videos based on user behavior data of different explanatory scenarios, wherein user behavior data is used to represent the behavior data generated by participating users in different explanatory scenarios; the association module is used to associate the target explanatory video with the description page of the first object, wherein the description page is used to provide descriptive information of the first object.

[0148] The acquisition module 502, determination module 504, and association module 506 described above correspond to steps S202 to S206 in the above embodiments. The three modules and their corresponding steps implement the same instances and application scenarios, but are not limited to the content disclosed in the above embodiments. The modules or units described above may be hardware or software components stored in memory and processed by one or more processors. These modules may also be part of a device and may run in the server 10 provided in the above embodiments.

[0149] In this embodiment of the application, the determining module is used to use a scoring model to score multiple explanation videos based on user behavior data of different explanation scenarios, and obtain score values ​​for multiple explanation videos. The score values ​​are used to quantify the explanation quality of the first object in different explanation videos. Based on the score values ​​of multiple explanation videos, the multiple explanation videos are filtered to obtain the target explanation video.

[0150] In this embodiment of the application, the determining module is used to extract features from user behavior data on multiple feature extraction dimensions to obtain multiple user behavior features; based on the feature weights corresponding to the multiple feature extraction dimensions and the multiple user behavior features, the target features for different explanation scenarios are determined; based on the target features for different explanation scenarios, the rating values ​​for multiple explanation videos are determined.

[0151] In this embodiment of the application, the determining module is used to map the target features of different explanation scenarios based on a preset correspondence relationship to obtain the rating values ​​of multiple explanation videos. The preset correspondence relationship is used to represent the correspondence between different features and different rating values.

[0152] In this embodiment of the application, the determining module is used to determine the explanation video with the highest score among the multiple explanation videos as the target explanation video based on the score values ​​of the multiple explanation videos.

[0153] In this embodiment of the application, the device is used to output multiple explanatory videos and multiple rating values ​​for the explanatory videos; in response to receiving feedback information on the rating values ​​of the multiple explanatory videos, the device updates the model parameters of the rating model based on the feedback information, wherein the feedback information is obtained by adjusting the rating values ​​of the multiple explanatory videos according to the multiple explanatory videos.

[0154] In this embodiment of the application, the acquisition module is used to acquire multiple target videos, wherein the different target videos are videos in which the second object gives a presentation in different presentation scenarios; content detection is performed on the multiple target videos to obtain the presentation time period of the first object; based on the presentation time period, the multiple target presentation videos are sliced ​​to obtain multiple presentation videos.

[0155] In this embodiment of the application, the device is used to output a target explanation video to the client of the second object when the first object has a new explanation scenario and the second object is explaining the first object.

[0156] According to an embodiment of this application, a schematic diagram of a video processing apparatus for implementing the above-described video processing method is also provided. Figure 6 This is a schematic diagram of a video processing apparatus according to an embodiment of this application, such as... Figure 6 As shown, the device includes: an acquisition module 602, a determination module 604, an association module 606, and an output module 608.

[0157] The acquisition module is used to acquire multiple explanatory videos of a first object by calling a first interface, wherein different explanatory videos are videos that explain the first object in different explanatory scenarios, and the first interface includes a first parameter whose value includes the multiple explanatory videos; the determination module is used to determine a target explanatory video from the multiple explanatory videos based on user behavior data of different explanatory scenarios, wherein the user behavior data represents the behavior data generated by participating users in different explanatory scenarios; the association module is used to associate the target explanatory video with the description page of the first object, wherein the description page provides descriptive information of the first object; and the output module is used to output the target explanatory video by calling a second interface, wherein the second interface includes a second parameter whose value includes the target explanatory video.

[0158] The aforementioned acquisition module 602, determination module 604, association module 606, and output module 608 correspond to steps S402 to S408 in the above embodiments. The four modules and their corresponding steps implement the same instances and application scenarios, but are not limited to the content disclosed in the above embodiments. The aforementioned modules or units may be hardware or software components stored in memory and processed by one or more processors. These modules may also be part of a device and may run in the server 10 provided in the above embodiments.

[0159] The preferred embodiments involved in the above embodiments of this application are the same as the solutions, application scenarios and implementation processes provided in the above embodiments, and will not be repeated here.

[0160] Embodiments of this application may provide a computing device. Figure 7 This is a structural block diagram of a computing device according to an embodiment of this application. Figure 7 As shown, the computing device 100 may include one or more (one shown in the figure) processors 102, memory 104, memory controller, and peripheral interfaces.

[0161] The aforementioned computing device can be understood as an integrated intelligent terminal, including but not limited to servers, desktop computers, PCs (Personal Computers), and all-in-one model machines. Furthermore, the computing device may have the model described in the above embodiments of this application pre-installed.

[0162] Specifically, this computing device can pre-install various types of models, including but not limited to models in fields such as natural language processing, visual processing, speech processing, code processing, and multimodal task processing, thus providing diverse model choices. In different product forms, this computing device can support one or more model usage methods, including but not limited to model training, model invocation, model fine-tuning, model deployment, model inference, and application. In some product forms, this computing device also supports model management, including but not limited to multi-type model management (supporting the management of discriminative, generative, and other model types), model version control (supporting the control of different model versions), and model evaluation (evaluating model performance and effectiveness based on model evaluation tools). In other product forms, this computing device can also create applications based on models, providing API calling capabilities. Models can be called into created applications through API interfaces, and application management tools are provided to control and manage applications.

[0163] Furthermore, this computing device can also include data management (supporting the creation and management of model tuning datasets), a training center (providing abundant training resources to help users learn and master AI technologies), and basic control capabilities (providing enterprise-level basic control capabilities to ensure system security and efficient operation). Through these functions, it provides a comprehensive, integrated device for AI development, training, deployment, and application.

[0164] The memory can be used to store software programs and modules, such as the program instructions / modules corresponding to the methods and apparatus in the embodiments of this application. The processor executes various functional applications and data processing by running the software programs and modules stored in the memory, thereby implementing the methods in the above embodiments. The memory may include high-speed random access memory (RAM) and non-volatile memory, such as one or more magnetic storage devices, flash memory, or other non-volatile solid-state memories. In some instances, the memory may further include memory remotely located relative to the processor, and these remote memories can be connected to terminal A via a network. Examples of such networks include, but are not limited to, the Internet, corporate intranets, local area networks (LANs), mobile communication networks, and combinations thereof.

[0165] The processor can invoke an executable program stored in memory via a transmission device to execute the method described in any of the above embodiments.

[0166] Embodiments of this application may provide an electronic device. Figure 8 This is a structural block diagram of an electronic device according to an embodiment of this application. Figure 8 As shown, the electronic device may include: an input / output device 112; a memory 114; and a processor 116, wherein the processor 116 is connected to the input / output device 112 and the memory 114 via a bus 118.

[0167] The memory can be used to store software programs and modules, such as the program instructions / modules corresponding to the methods and apparatus in the embodiments of this application. The processor executes various functional applications and data processing by running the software programs and modules stored in the memory, thereby implementing the methods in the above embodiments. The memory may include high-speed random access memory (RAM) and non-volatile memory, such as one or more magnetic storage devices, flash memory, or other non-volatile solid-state memories. In some instances, the memory may further include memory remotely located relative to the processor, and these remote memories can be connected to terminal A via a network. Examples of such networks include, but are not limited to, the Internet, corporate intranets, local area networks (LANs), mobile communication networks, and combinations thereof.

[0168] The processor can invoke an executable program stored in memory via a transmission device to execute the method described in any of the above embodiments.

[0169] Those skilled in the art will understand that, Figure 8The structure shown is schematic. Electronic devices can also be smartphones (such as Android phones, iOS phones, etc.), tablets, PDAs, mobile internet devices (MIDs), PADs, and other terminal devices. This diagram does not limit the structure of the aforementioned electronic devices. For example, electronic devices may include more or fewer components (such as network interfaces, display devices, etc.) than shown in the diagram, or have different configurations than those shown.

[0170] Those skilled in the art will understand that all or part of the steps in the various methods of the above embodiments can be implemented by a program instructing the hardware related to the terminal device. The program can be stored in a computer-readable storage medium, which may include: a flash drive, read-only memory (ROM), random access memory (RAM), a magnetic disk, or an optical disk, etc.

[0171] Embodiments of this application also provide a computer-readable storage medium. Optionally, in this embodiment, the aforementioned computer-readable storage medium can be used to store program code executed by the method provided in the above embodiments.

[0172] Optionally, in this embodiment, the storage medium may be located in a computing device or an electronic device.

[0173] Optionally, in this embodiment, the computer-readable storage medium is configured to store an executable program. When the executable program runs, it controls the device where the computer-readable storage medium is located to perform the method described in any of the above embodiments.

[0174] Embodiments of this application also provide a computer program product. Optionally, in this embodiment, the computer program product may include a computer program. When executed by a processor, the computer program implements the methods provided in the above embodiments.

[0175] Embodiments of this application also provide a computer program product. Optionally, the computer program product may include a non-volatile computer-readable storage medium. The non-volatile computer-readable storage medium can be used to store a computer program. When the computer program is executed by a processor, it implements the method provided in the above embodiments.

[0176] Embodiments of this application also provide a computer program. Optionally, in this embodiment, when the computer program is executed by a processor, it implements the method provided in the above embodiments.

[0177] In the above embodiments of this application, the descriptions of each embodiment have different focuses. For parts not described in detail in a certain embodiment, please refer to the relevant descriptions of other embodiments.

[0178] In the several embodiments provided in this application, it should be understood that the disclosed technical content can be implemented in other ways. The device embodiments described above are illustrative; for example, the division of units is a logical functional division, and in actual implementation, there may be other division methods. For example, multiple units or components may be combined, integrated into another system, or some features may be ignored or not executed. Furthermore, the coupling, direct coupling, or communication connection shown or discussed may be through some interfaces, indirect coupling of units or modules, or communication connection, and may be electrical or other forms.

[0179] The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the units can be selected according to actual needs to achieve the purpose of this embodiment.

[0180] Furthermore, the functional units in the various embodiments of this application can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit. The integrated unit can be implemented in hardware or as a software functional unit.

[0181] If the integrated unit is implemented as a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of this application, in essence, or the part that contributes to the prior art, or all or part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of this application. The aforementioned storage medium includes various media capable of storing program code, such as a USB flash drive, read-only memory (ROM), random access memory (RAM), portable hard drive, magnetic disk, or optical disk.

[0182] The above description represents the preferred embodiments of this application. For those skilled in the art, various improvements and modifications can be made without departing from the principles of this application, and these improvements and modifications should also be considered within the scope of protection of this application.

Claims

1. A method of processing a video, characterized by, The method comprises: obtaining a plurality of explanation videos of a first object, wherein different explanation videos are videos for explaining the first object in different explanation scenes; determining a target explanation video from the plurality of explanation videos based on user behavior data of different explanation scenes, wherein the user behavior data is used to represent behavior data generated by participating users in different explanation scenes; associating the target explanation video to a description page of the first object, wherein the description page is used to provide description information of the first object.

2. The method of claim 1, wherein, Determining a target explanation video from the plurality of explanation videos based on user behavior data of different explanation scenes comprises: using a scoring model to score the plurality of explanation videos based on the user behavior data of different explanation scenes to obtain scoring values of the plurality of explanation videos, wherein the scoring values are used to quantify the explanation quality of the first object in different explanation videos; filtering the plurality of explanation videos based on the scoring values of the plurality of explanation videos to obtain the target explanation video.

3. The method of claim 2, wherein, Using a scoring model to score the plurality of explanation videos based on the user behavior data of different explanation scenes to obtain scoring values of the plurality of explanation videos comprises: performing feature extraction on the user behavior data in a plurality of feature extraction dimensions to obtain a plurality of user behavior features; determining target features of different explanation scenes based on feature weights corresponding to the plurality of feature extraction dimensions and the plurality of user behavior features; determining scoring values of the plurality of explanation videos based on the target features of different explanation scenes.

4. The method of claim 3, wherein, Determining scoring values of the plurality of explanation videos based on the target features of different explanation scenes comprises: mapping the target features of different explanation scenes based on a preset correspondence relationship to obtain the scoring values of the plurality of explanation videos, wherein the preset correspondence relationship is used to represent the correspondence between different features and different scoring values.

5. The method of claim 2, wherein, Filtering the plurality of explanation videos based on the scoring values of the plurality of explanation videos to obtain the target explanation video comprises: determining, based on the scoring values of the plurality of explanation videos, that an explanation video with the largest scoring value in the plurality of explanation videos is the target explanation video.

6. The method of claim 2, wherein, The method further comprises: outputting the plurality of explanation videos and the scoring values of the plurality of explanation videos; updating model parameters of the scoring model based on feedback information of the scoring values of the plurality of explanation videos in response to receiving the feedback information, wherein the feedback information is obtained by adjusting the scoring values of the plurality of explanation videos based on the plurality of explanation videos.

7. The method of claim 1, wherein, Obtaining a plurality of explanation videos of the first object comprises: obtaining a plurality of target videos, wherein different target videos are videos for explaining a second object in different explanation scenes; performing content detection on the plurality of target videos to obtain an explanation time period of the first object; based on the explanation time period, respectively slicing the plurality of target videos to obtain the plurality of explanation videos.

8. The method of claim 7, wherein, The method further comprises: If a new explanation scene of the first object exists, output the target explanation video in a client of the second object in an explanation process of the second object for the first object.

9. A method of processing a video, characterized by, The method comprises: obtaining a plurality of explanation videos of a first object by calling a first interface, wherein different explanation videos are videos of the first object explained in different explanation scenes, the first interface comprises a first parameter, and a parameter value of the first parameter comprises the plurality of explanation videos; determining a target explanation video from the plurality of explanation videos based on user behavior data of different explanation scenes, wherein the user behavior data is used to represent behavior data generated by participating users in different explanation scenes; associating the target explanation video to a description page of the first object, wherein the description page is used to provide description information of the first object; outputting the target explanation video by calling a second interface, wherein the second interface comprises a second parameter, and a parameter value of the second parameter comprises the target explanation video.

10. An electronic device, comprising: The method comprises: a memory storing an executable program; a processor connected with the memory through a bus, used to run the program, wherein the program performs the method of any one of claims 1 to 9 when running.

11. A computer readable storage medium, characterized in that, The computer readable storage medium comprises a stored executable program, wherein the computer readable storage medium controls a device where the computer readable storage medium is located to perform the method of any one of claims 1 to 9 when the executable program runs.

12. A computer program product, characterised in that, The computer program is executed by the processor to implement the method of any one of claims 1 to 9.

Images