Unverified50% confidenceFactExact time
D-ID、HeyGen等数字人视频生成平台主要聚焦于面部运动迁移，通过音频驱动的面部关键点预测实现唇型同步和面部表情动画
1
Sources
50%
Confidence
Medium-term (~90 days)
Relevance
8/24/2026
First Seen
Valid until: 11/22/2026
Sources
Related Entities
Related Claims
Unverified当前许多平台的人脸识别系统仍然依赖2D视频验证，容易被AI换脸技术突破70% similarUnverifiedAsking the other party to wave their hand in front of their face or turn to show a profile view during a video call can be used to identify AI face-swap forgeries.70% similarUnverified随着FaceSwap、DeepFaceLab等开源项目的成熟以及基于Stable Diffusion的换脸工具涌现，部分工具仅需一张照片即可生成可信度较高的动态换脸视频69% similarVerifiedHugging Face Transformers covers machine learning models across multiple modalities including text, vision, audio, and multimodal.69% similarUnverifiedDeep learning-based avatar driving solutions include models such as Audio2Face and SadTalker, which can directly generate facial animations from audio.67% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/795370API
curl https://kongchang.com/api/v1/knowledge/claims/795370MCP
get_claim(id=795370)