基于表情捕捉的数字人面部绑定生成方法
By real-time acquisition and optimization of digital human facial binding methods, the problems of low binding efficiency and unnatural expressions in existing technologies have been solved, achieving efficient and natural digital human facial movements, which are suitable for film and television, virtual live streaming and interactive games.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- ZHEJIANG VERSATILE MEDIA
- Filing Date
- 2025-09-26
- Publication Date
- 2026-07-17
AI Technical Summary
Existing digital human facial binding methods rely on manual adjustments or preset templates, resulting in low binding efficiency, poor flexibility, inability to capture subtle changes in facial expressions, and stiff and unnatural expressions, making it difficult to meet the requirements for high realism.
Facial data is collected in real time by facial expression capture devices, facial expression feature vectors are extracted, facial motion trajectory fitting and binding rule generation are performed, and the topological structure of the digital human model is combined for mapping and optimization to generate optimized facial binding data to drive the digital human's facial movements.
It achieves delicate and natural facial expressions in digital humans, improves binding efficiency and adaptability, enhances the performance of digital humans in film, virtual live streaming and interactive games, and reduces manual operation costs.
Smart Images
Figure CN121214518B_ABST