CN-121985191-A - Video generation method and system, video center station, medium and program product

Abstract

The specification provides a video generation method and system, a video middle station, a medium and a program product, wherein the method comprises the steps of receiving a video generation request sent by a seller client, responding to the video generation request, acquiring commodity metadata and commodity material sets of target commodities from a data server, determining target commodity materials used for generating videos from the commodity material sets based on the commodity metadata of the target commodities, and inputting the target commodity materials into a video generation model so that the video generation model generates target videos based on the target commodity materials.

Inventors

• Tang Nengfa
• ZHANG JIALI
• YIN LIMIN

Assignees

• 杭州阿里巴巴海外互联网产业有限公司
• 阿里巴巴新加坡控股有限公司

Dates

Publication Date

20260505

Application Date

20251215

Claims (17)

1. 1. A video generation method is applied to a video middle station connected with a data server of an electronic commerce platform and a seller client of the electronic commerce platform, wherein the data server is used for storing commodity metadata and commodity materials related to commodities sold on the electronic commerce platform, and the method comprises the following steps: receiving a video generation request sent by the seller client; In response to the video generation request, acquiring commodity metadata and commodity material sets of a target commodity from the data server, and determining target commodity materials for generating videos from the commodity material sets based on the commodity metadata of the target commodity; and inputting the target commodity materials into a video generation model so that the video generation model generates target videos based on the target commodity materials.
2. 2. The method of claim 1, the determining target merchandise material for generating video from the set of merchandise materials based on merchandise metadata of the target merchandise, comprising: the method comprises the steps of acquiring pre-generated thinking chain prompt information, wherein the thinking chain prompt information is used for defining a multi-level reasoning step of screening commodity materials based on commodity metadata, and the output of each level of reasoning step is used as the input of at least one level of subsequent reasoning step; And carrying out thinking chain reasoning on the commodity metadata of the target commodity and the commodity material set according to the multistage reasoning step by using the thinking chain prompt information, the commodity metadata of the target commodity and the commodity material set target reasoning model, and determining the target commodity material according to a reasoning result.
3. 3. The method of claim 2, the step of multi-level reasoning comprising: Acquiring task background information of a video generation task, wherein the task background information comprises screening standards of commodity materials, analysis flows of the commodity materials, first evaluation standards for evaluating whether the commodity materials are matched with commodity metadata, user information of video audience users and second evaluation standards for evaluating whether the commodity materials are matched with the video audience users; Screening a plurality of candidate commodity materials meeting the screening standard from the commodity material set; Analyzing each screened candidate commodity material according to the analysis flow to determine material information of the candidate commodity materials; Evaluating a first degree of matching between the material information of the candidate commodity material and commodity metadata of the target commodity based on the first evaluation criterion; evaluating a second degree of matching between the material information of the candidate commodity material and the user information of the video audience user based on the second evaluation criterion; determining the weight of the candidate commodity material based on the first matching degree and the second matching degree; and screening target commodity materials from the plurality of candidate commodity materials based on the weights of the plurality of candidate commodity materials.
4. 4. The method of claim 2, the method further comprising: acquiring sample data, wherein the sample data comprises sample thinking chain prompt information, sample commodity metadata, sample commodity materials and label information, the sample thinking chain prompt information comprises a plurality of levels of reasoning steps and label output results corresponding to each level of reasoning steps, and the label information is used for indicating whether the sample commodity materials are screened as materials for generating videos; Inputting the sample data into a base model, so that the base model carries out thinking chain reasoning on the sample commodity metadata and the sample commodity materials based on multistage reasoning steps in the sample thinking chain prompt information to obtain reasoning results corresponding to each stage of reasoning steps; determining the loss corresponding to each level of reasoning step based on the difference between the reasoning result respectively corresponding to each level of reasoning step and the labeling output result corresponding to each level of reasoning step output by the base model; determining total loss based on the losses corresponding to each level of reasoning steps; And training the base model based on the total loss to obtain the target reasoning model.
5. 5. The method of claim 1, wherein the video generation request carries seller identification information, the seller identification information being used for representing a target seller associated with the seller client, and wherein the obtaining commodity metadata and commodity material sets of a target commodity from the data server in response to the video generation request comprises: Acquiring a commodity list published by the target seller on the electronic commerce platform from the data server based on the seller identification information, and returning the commodity list to the seller client; And in response to the seller client selecting the target commodity from the commodity list, acquiring commodity metadata and commodity material sets of the target commodity from the data server.
6. 6. The method of claim 1, after determining target merchandise material for generating video from the set of merchandise materials based on merchandise metadata of the target merchandise, the method further comprising: Returning the screening result of the commodity material set and the target commodity material to the seller client for display; And responding to an adjustment instruction of the seller client side to the target commodity material, and adjusting the target commodity material.
7. 7. The method of claim 1, the inputting the target commodity material into a video generation model to cause the video generation model to generate video based on the target commodity material, comprising: generating a video abstract based on commodity metadata of the target commodity; and inputting the video abstract and the target commodity material into a video generation model so that the video generation model generates target video based on the video abstract and the target commodity material.
8. 8. The method of claim 7, prior to inputting the video summary and the target commodity material into a video generation model, the method further comprising: returning the video summary to the seller client; and modifying the video abstract in response to receiving a modification instruction of the seller client.
9. 9. The method of claim 1, the inputting the target commodity material into a video generation model to cause the video generation model to generate a target video based on the target commodity material, comprising: Inputting the target commodity materials into a video generation model so that the video generation model generates a plurality of video preview files based on the target commodity materials; Sending the plurality of video preview files to the seller client for display; And generating a target video with visual characteristics corresponding to the target video preview file in response to receiving the target video preview file selected by the seller client from the plurality of video preview files.
10. 10. The method of claim 1, the inputting the target commodity material into a video generation model to cause the video generation model to generate a target video based on the target commodity material, comprising: Acquiring video attribute information sent by the seller client; and inputting the video attribute information and the target commodity material into a video generation model so that the video generation model generates target video with the video attribute information based on the target commodity material.
11. 11. The method of claim 1, wherein the video center supports a plurality of video generation modes, wherein different video generation modes are used for generating different types of target videos, and wherein different types of target videos are generated by different video generation models, wherein the inputting the target commodity material into a video generation model to cause the video generation model to generate a target video based on the target commodity material comprises: acquiring a target video generation mode selected by the seller client from the plurality of video generation modes; And inputting the target commodity materials into a video generation model corresponding to the target video generation mode, so that the video generation model corresponding to the target video generation mode generates target videos based on the target commodity materials.
12. 12. The method of claim 1, the method further comprising: translating the target video, and/or Video fission of the target video, and/or Publishing the target video to a designated platform, and/or Video transcoding the target video, and/or And detecting the risk of the target video.
13. 13. A video generation method is applied to a seller client side of an e-commerce platform, the seller client side is in communication connection with a video center, the video center is also in communication connection with a data server of the e-commerce platform, the data server is used for storing commodity metadata and commodity materials related to commodities sold on the e-commerce platform, and the method comprises the following steps: Displaying a video generation control on a seller background management page; In response to detecting an operation instruction for the video generation control, sending a video generation request to the video center station; The video generation model called by the video center station generates the target video based on target commodity materials, the target commodity materials are determined from commodity material sets of target commodities based on commodity metadata of the target commodities, and commodity metadata and commodity material sets of the target commodities are obtained from a data server of the electronic commerce platform.
14. 14. A video midstand comprising a processor, a memory for storing processor executable instructions, wherein the processor is configured to implement the steps of the method according to any of claims 1-12 by executing the executable instructions.
15. 15. A video generation system comprises an e-commerce platform, a seller client and a video generation middle station; the electronic commerce platform comprises a data server, wherein the data server is used for storing commodity metadata and commodity materials related to commodities sold on the electronic commerce platform; the seller client is used for sending a video generation request to the video middle station; the video midstand is adapted to perform the method according to any of claims 1-12.
16. 16. A computer readable storage medium having stored thereon computer instructions which, when executed by a processor, implement the steps of the method of any of claims 1-13.
17. 17. A computer program product comprising computer programs/instructions which, when executed by a processor, implement the steps of the method of any of claims 1-13.

Description

Video generation method and system, video center station, medium and program product Technical Field One or more embodiments of the present disclosure relate to the field of video generation technology, and in particular, to a video generation method and system, a video center station, a medium, and a program product. Background In current ecosystems of electronic commerce, video content has become a key medium for attracting consumers, displaying merchandise details, and improving conversion rates. For the sellers of the vast electronic commerce, the high-quality commodity video is manufactured efficiently and at low cost, so that the urgent requirements of enhancing the competitive power of shops and adapting to the trend of content marketing are met. Currently, the dominant way sellers generate commodity videos typically relies on manual operations and third party tools. The seller needs to collect or shoot the original materials such as commodity pictures, video clips, texts, music and the like by himself and upload the materials to independent third-party video editing software or an online production platform for video production. However, on one hand, the quantity of materials uploaded by the seller is often relatively small, which results in poor quality and richness of the generated video, and on the other hand, the seller needs to execute a series of operations such as material collection and uploading in the video generation process, which results in complex operation process and low video generation efficiency. Disclosure of Invention In view of this, one or more embodiments of the present disclosure provide the following technical solutions: According to a first aspect of one or more embodiments of the present specification, a video generating method is provided, which is applied to a video middle station connected with a data server of an e-commerce platform and a seller client of the e-commerce platform, wherein the data server is used for storing commodity metadata and commodity materials related to commodities sold on the e-commerce platform, and the method comprises: receiving a video generation request sent by the seller client; In response to the video generation request, acquiring commodity metadata and commodity material sets of a target commodity from the data server, and determining target commodity materials for generating videos from the commodity material sets based on the commodity metadata of the target commodity; and inputting the target commodity materials into a video generation model so that the video generation model generates target videos based on the target commodity materials. According to a second aspect of one or more embodiments of the present specification, there is provided a video generating method applied to a seller client of an e-commerce platform, the seller client being communicatively connected to a video middle station, the video middle station being also communicatively connected to a data server of the e-commerce platform, the data server being configured to store commodity metadata and commodity materials related to a commodity sold on the e-commerce platform, the method comprising: Displaying a video generation control on a seller background management page; In response to detecting an operation instruction for the video generation control, sending a video generation request to the video center station; The video generation model called by the video center station generates the target video based on the target commodity material, the target commodity material is determined from a commodity material set of the target commodity based on commodity metadata of the target commodity, and the commodity metadata and the commodity material set of the target commodity are obtained from a data server of the electronic commerce platform. According to a third aspect of one or more embodiments of the present specification, there is provided a video midstand comprising a processor, a memory for storing processor executable instructions, wherein the processor is configured to implement the steps of the method according to the first aspect of one or more embodiments of the present specification by executing the executable instructions. According to a fourth aspect of one or more embodiments of the present specification, a video generation system is presented, comprising an e-commerce platform, a vendor client, and a video generation middle station; the electronic commerce platform comprises a data server, wherein the data server is used for storing commodity metadata and commodity materials related to commodities sold on the electronic commerce platform; the seller client is used for sending a video generation request to the video middle station; The video center station is configured to perform the method according to the first aspect of one or more embodiments of the present specification. According to a fifth aspect of one or more embodiments of the present descripti