Method, apparatus and electronic device for generating virtual image

By integrating local and global facial feature information into a 3D face generation model and combining it with a self-attention mechanism to process images, the problem of low realism of virtual images in the metaverse is solved, achieving realistic 3D face model generation and improving user experience.

CN116524564BActive Publication Date: 2026-07-21INDUSTRIAL AND COMMERCIAL BANK OF CHINA
View PDF 2 Cites 0 Cited by

Patent Information

Authority / Receiving Office
CN · China
Patent Type
Patents(China)
Current Assignee / Owner
INDUSTRIAL AND COMMERCIAL BANK OF CHINA
Filing Date
2023-04-28
Publication Date
2026-07-21

AI Technical Summary

Technical Problem

Existing technologies generate virtual images of customers in the metaverse with low realism, and traditional face reconstruction processes do not fully consider lighting and facial expression factors, resulting in poor quality of reconstructed 3D face models.

Method used

A 3D face generation model is adopted. By fusing local and global face feature information and combining a self-attention mechanism, illumination feature information is extracted from the target image. Multiple convolutional and pooling layers are used to process the image, generate a 3D face feature vector, and generate a virtual image in the metaverse.

Benefits of technology

It improves the realism of the virtual avatar of the customer in the metaverse by extracting and fusing local and global facial feature information to generate a realistic 3D facial model, thereby enhancing the user experience.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN116524564B_ABST
    Figure CN116524564B_ABST
Patent Text Reader

Abstract

The application discloses a virtual image generation method and device and electronic equipment. The method is applied to the field of artificial intelligence, and comprises the following steps: obtaining a target image of a target customer; processing the target image by using a three-dimensional face generation model to obtain a three-dimensional face feature vector corresponding to the target image; determining a three-dimensional face coordinate corresponding to the target customer according to the three-dimensional face feature vector; and generating a virtual image corresponding to the target customer in the metaverse according to the three-dimensional face coordinate. Through the application, the problem that the virtual image corresponding to the customer in the metaverse is low in reality in the related art is solved.
Need to check novelty before this filing date? Find Prior Art