Incredibly realistic human videos from just a single image and a motion signal
- Istanbul Digital Economy Club
- Feb 6
- 1 min read
Deepseek is yesterday's China news...
Bytedance, Tiktok's mother company, shared demos of OmniHuman this week, an AI framework that generates realistic human videos from a single image and a motion signal, like audio or video. It uses multimodal motion conditioning to translate these inputs into lifelike movements, gestures, and expressions.
OmniHuman works with portraits, half-body, and full-body images. It can also animate non-human subjects, such as cartoons or animals, making it highly versatile.
The AI model is currently in the research phase. The developers have shared demos and hinted at a code release in the future, but it’s not accessible to the public at this time.
See links to more OmniHuman examples and their introduction in comments.
Comentários