Digital human live broadcast goods carrying method based on ChatGPT

By generating answers in ChatGPT and using digital human voice model to display answers, the problem of insufficient human resources and interactive capabilities for the human settings of digital human live broadcast application is solved, and an efficient and intelligent live broadcasting mode is realized, satisfying the live broadcast business diversified needs.

CN118921533APending Publication Date: 2024-11-08SUZHOU GUOBENXIN INFORMATION TECH CO LTD
View PDF 0 Cites 1 Cited by

Patent Information

Application Number
CN202410220434.6
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2024-08-26
Publication Date
2024-11-08

AI Technical Summary

Technical Problem

The existing digital live broadcast application artificially sets the explanation content to consume a lot of human resources, the content is monotonous, and the interactive ability is weak, which cannot meet the needs of one Q&A live broadcast business.

Method used

By generating answers in ChatGPT, using AI models to automatically generate introduction content and real-time answer barrage, reducing human resource consumption, real-time question answering, and combining digital human voice model to display answers.

Benefits of technology

It simplifies the operation process, reduces development costs, improves user experience and live broadcast mode intelligence, reduces the burden on live broadcasters, and realizes the integration and interactivity of functions.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN118921533A_ABST
    Figure CN118921533A_ABST
Patent Text Reader

Abstract

The invention relates to a digital human live broadcast goods carrying method based on ChatGPT. The method comprises the following steps: S1, establishing and initializing digital human and live broadcast information; s2, inputting basic commodity information; s3, generating introduction statements through ChatGPT, and sequentially displaying the introduction statements; and S4, ending after a set requirement is met, or capturing bullet screen information of the live broadcast room. According to the technical scheme provided by the invention, AI automatic generation is introduced into a manual input part, so that the operation is greatly simple and convenient, and the human resource consumption is reduced; a function integrated interface facilitates unified management of a user; the process is simple and clear, each function is independently realized, the information processing and data returning speed is greatly increased, and the manufacturing cost of developers is reduced; chatGPT is introduced for interaction, the application function is more colorful and intelligent, and the user experience is better; the live broadcast mode is intelligent, and the auxiliary live broadcast mode carries out live broadcast answering through manpower and application, thereby reducing the burden of a live broadcaster, and enabling the live broadcaster to complement each other to achieve the maximum value.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The present invention relates to the technical field of live streaming and selling goods, and specifically to a digital human live streaming and selling goods method based on ChatGPT. Background Art

[0002] With the strong explosion of AI technology in recent years, digital humans can be completed faster and more efficiently in terms of production costs, generation methods, rendering processes, and rapid deployment. This has also expanded AI digital humans' diversified scene application capabilities, especially in the field of live broadcasts and short videos. AI digital humans have brought new growth momentum to the live broadcast industry with low-input, high-output, and long-lasting live broadcast modes. Customizing an AI digital human virtual anchor through the digital human live broadcast system only requires a small amount of money per month to support the most basic exposure needs of the brand's live broadcast room. Merchants no longer need to spend a lot of money to hire anchors and customer service personnel, nor do they need to arrange live broadcast venues and layout scenes. Live broadcasts and online explanations can be achieved through AI digital humans.

[0003] Relying on AI's multimodal interaction and generation capabilities, companies only need to enter text into the system to automatically generate a digital human virtual anchor's oral video. In addition, through model training, the matching degree between the digital population type and the content is ensured, and high-fidelity sound materials can be obtained, making the virtual anchor more natural in terms of expression, voice and language generation, just like a real person. AI digital humans greatly save the time and cost of real anchors, and no longer need to consider various factors such as manpower, clothing, makeup, venue, time, etc. At the same time, it improves the efficiency of making videos, allowing the creative team to focus more on content planning and production.

[0004] However, general corporate digital human live broadcast applications require a lot of manually set explanations, consume a certain amount of human resources, and the content is relatively monotonous compared to traditional anchors. In addition, as a live broadcast application, there is little communication content with the purchaser, that is, the audience, and the interactive ability is weak, which cannot meet the basic live broadcast business needs of question and answer. Summary of the invention

[0005] 1. Technical issues to be resolved

[0006] In order to solve the above technical problems, the present invention provides a method for live streaming with digital humans based on ChatGPT. When users set product parameters, the present invention allows users to input only the most basic information, and the AI ​​model can automatically generate introduction content, greatly reducing the consumption of human resources. In addition, after collecting and crawling the bullet screen content sent by the audience in the live broadcast room, the answer is generated in ChatGPT, and the content is returned to the digital human, and then displayed through the digital human voice model, so as to answer real-time questions and meet various needs of the live broadcast business.

[0007] (II) Technical solution

[0008] Based on this, the present invention provides the following technical solution: a digital human live streaming method based on ChatGPT, comprising the following steps:

[0009] Step S1: Create and initialize digital human and live broadcast information;

[0010] Step S2: Enter basic product information;

[0011] Step S3: Generate introduction sentences through ChatGPT and display them sequentially;

[0012] Step S4: End after the setting requirements are met, or capture the live broadcast room barrage information; classify the barrage information and identify question barrage and other barrage;

[0013] When it is a question barrage, the answer is generated through the ChatGPT training model and displayed through audio; then the operation of grabbing the barrage information in the live broadcast room is repeated; finally, the question and answer content are collected into the cloud server database;

[0014] When it is other barrage, answer the set sentence, guide the audience to ask questions, and then repeat the operation of capturing the barrage information in the live broadcast room.

[0015] (III) Beneficial effects

[0016] Compared with the prior art, the present invention provides a digital human live streaming method based on ChatGPT, which has the following beneficial effects:

[0017] The digital human live streaming method for selling goods based on ChatGPT introduces AI automatic generation into the manual input part of the technical solution proposed in the present invention, which greatly simplifies the operation and reduces the consumption of human resources; the function-integrated interface facilitates unified management by users; the process is simple and clear, and each function is independently implemented, which greatly increases the speed of processing information and returning data, and reduces the manufacturing cost of developers; ChatGPT is introduced for interaction, and the application functions are more colorful and intelligent, so that the user experience is better; the live broadcast mode is intelligent, and the auxiliary live broadcast mode uses manpower and applications to jointly broadcast answers, reducing the burden on the live broadcaster, and complementing each other to maximize the value. BRIEF DESCRIPTION OF THE DRAWINGS

[0018] Figure 1 It is a flow chart of the processing method of the present invention;

[0019] Figure 2 This is a flow chart of the auxiliary live broadcast mode processing method of the present invention. DETAILED DESCRIPTION

[0020] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0021] Example 1

[0022] See also Figure 1 A method for bringing goods through live broadcasting with digital humans based on ChatGPT. When users set product parameters, the present invention allows users to enter only the most basic information, and the AI ​​model can automatically generate introduction content, greatly reducing the consumption of human resources. In addition, after collecting and crawling the bullet screen content sent by the audience in the live broadcast room, the answer is generated in ChatGPT, and the content is returned to the digital human, and then displayed through the digital human voice model, so as to answer real-time questions and meet various needs of the live broadcast business.

[0023] The specific steps are as follows:

[0024] First: Manually input basic parameters of the products being sold and upload other related content.

[0025] Second: After the application collects input, it automatically generates product introduction sentences through the training model to make them coherent and complete, and the content of the introduction generated each time will be different to ensure vividness.

[0026] Third: Generate the introduction sentence into MP3 format audio through the digital human voice model and output it to the live broadcast room.

[0027] Fourth: Crawl the barrage information sent by the audience in the live broadcast room and return it to the application.

[0028] Fifth: Pass the question and generate answers to the relevant comments through the trained ChatGPT sales model.

[0029] Sixth: Convert the answer sentence into MP3 format audio through the digital human voice model and output it to the live broadcast room.

[0030] Seventh: Delete temporary audio files, collect questions and answers, return them to the server database, and improve the digital human ChatGPT sales model through training again.

[0031] When users add product input information, they only need to fill in the data containing basic information and key information. The learned ChatGPT will automatically generate connecting words to make them coherent and become different product introductions. The major aspects include commodity categories, food categories, cultural travel categories, and cultural and sports categories. The key information includes introduction, advantages, parameters, instructions, etc.

[0032] If the app does not receive the live broadcast room bullet screen information within the specified time, the digital human will play the inputted pre-ordered product introductions in sequence. When receiving the bullet screen information, it will immediately start responding to the bullet screen after completing the current product introduction; if the live broadcast room bullet screen information is not received within the specified time again, the product introduction will continue to be played in the previous sequence.

[0033] After the application crawls the live broadcast room barrage information through the Python crawling module, it passes the questions in the barrage to the trained ChatGPT sales model to generate answers to the questions, generates a temporary MP3 file on the server, plays it in the live broadcast room with the help of the digital human, and deletes the temporary file after the playback is completed.

[0034] The server collects all the questions and generated answers into the database in the cloud server for subsequent reference and training of the ChatGPT sales model.

[0035] Example 2

[0036] See also Figure 2 , a method of selling goods through digital human live broadcast based on ChatGPT. Users can not only input information in the application, but also set the live broadcast mode. In addition to the normal live broadcast mode, during the peak period of the live broadcast room, users can also turn on the auxiliary live broadcast mode of the application, and conduct manual live broadcast during this period to improve the efficiency of selling goods. The application skips the step of generating voice answers and plays them, and instead returns the answers to the live broadcast room to answer some of the questions in the barrage in the form of barrage, reducing the burden on the live broadcaster. Traditional live broadcast and digital human live broadcast applications complement each other and play the greatest value together.

[0037] The present invention uses ChatGPT to interact with questions, generates interactions and answers quickly, and has intelligent and rich content, thereby meeting various needs of buyers.

[0038] The speech model part of the present invention generates a temporary MP3 file and plays it in the live broadcast room, which is deleted after the playing is completed, thereby improving the running speed of the generation and interaction process and greatly saving the memory and storage occupied by the digital human model speech module.

[0039] Although embodiments of the present invention have been shown and described, it will be appreciated by those skilled in the art that various changes, modifications, substitutions and variations may be made to the embodiments without departing from the principles and spirit of the present invention, and that the scope of the present invention is defined by the appended claims and their equivalents.

Claims

1. A digital human live streaming method based on ChatGPT, characterized by: The following steps are involved: Step S1: Create and initialize digital human and live broadcast information; Step S2: Enter basic product information; Step S3: Generate introduction sentences through ChatGPT and display them sequentially; Step S4: End after the setting requirements are met.

2. According to a ChatGPT-based digital human live streaming method according to claim 1, it is characterized by: The specific operations of step S4 above are as follows: Step S41: Capture live broadcast room bullet screen information; Step S42: Classify the barrage information.

3. A method for selling goods through digital human live streaming based on ChatGPT according to claim 2, characterized in that: The specific operations of step S42 are as follows: Step S421: Identify question comments and other comments.

4. According to a ChatGPT-based digital human live streaming method of claim 3, it is characterized by: The specific operations of step S421 are as follows: Step S42a: When it is a question barrage, the answer is generated through the ChatGPT training model and displayed through audio; Then repeat the operation of step S41; Finally, the questions and answers are collected and stored in the cloud server database.

5. According to a ChatGPT-based digital human live streaming method of claim 3, it is characterized by: The specific operations of step S421 are as follows: Step S42b: When it is other bullet screens, answer the set sentence, guide the audience to ask questions, and then repeat the operation of step S41.

Citation Information

Cited By

  • Synchronous live broadcast method for multiple live broadcast rooms based on AI digital human

    CN121531163A