A web-end digital human cross-form infringement detection and early warning method and system

By using a web-based digital human cross-form infringement detection and early warning method, the difficulties in cross-form comparison of digital humans and the problem of resisting skinning and plagiarism have been solved. It has achieved unified comparison of digital humans in multiple formats and efficient and accurate infringement detection, thereby improving privacy protection and the legal validity of detection results.

CN122618319BActive Publication Date: 2026-11-06CHINA ELECTRONICS STANDARDIZATION INST
View PDF 2 Cites 0 Cited by

Patent Information

Application Number
CN202610741524.9
Authority / Receiving Office
CN · China
Patent Type
Patents(China)
Current Assignee / Owner
Filing Date
2026-05-27
Publication Date
2026-11-06
Estimated Expiration
2046-05-27

AI Technical Summary

Technical Problem

Existing digital human infringement detection technologies suffer from problems such as difficulty in cross-morphological comparison, weak resistance to plagiarism and reskinning, heavy client dependence, contradictions between privacy protection and evidence preservation, and insufficient legal validity of detection results.

Method used

This paper presents a web-based method for cross-morphological infringement detection and early warning of digital humans. It achieves cross-morphological comparison and early warning by using multi-format adaptive preprocessing, cross-morphological standardized mapping, decoupling extraction of geometric bone structure and texture skin feature vectors, and a two-stage risk judgment strategy.

Benefits of technology

It achieves unified identity comparison of digital humans in multiple formats, including 2D images, 2.5D models, 3D models, and videos, overcoming the dimensional incompatibility and texture change sensitivity issues of traditional technologies. It supports static 3D files without skeleton binding and low-resolution input, reducing computational load and improving detection accuracy and privacy protection.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN122618319B_ABST
    Figure CN122618319B_ABST
Patent Text Reader

Abstract

This invention discloses a web-based method and system for cross-morphological infringement detection and early warning of digital humans, belonging to the field of digital content copyright protection technology. The method includes: receiving multi-format digital human files via a web interface and performing adaptive preprocessing; mapping different morphological data uniformly to a standard orthogonal view space; decoupling and extracting geometric skeletal and textural feature vectors; calculating the comprehensive similarity using an adaptive nonlinear fusion formula; determining the risk level through a two-stage graded early warning strategy; and performing graded privacy storage and judicial blockchain evidence preservation based on the risk level. This invention achieves unified infringement detection for 2D, 2.5D, and 3D digital humans, accurately identifying digital human structural plagiarism and counterfeiting behaviors such as skin replacement and texture theft. It requires no client installation, balances privacy protection and judicial evidence preservation needs, and effectively reduces the cost of digital human copyright protection and rights enforcement.
Need to check novelty before this filing date? Find Prior Art

Citation Information

Patent Citations

  • Image detection method and device, electronic equipment and readable storage medium

    CN112215812A

  • Digital human identification method and device and computer readable storage medium

    CN118537898A