ID-Animator
This repository is the official checkpoint of ID-Animator. It is a Zero-shot ID-Preserving Human Video Generation framework. It can generate high-quality ID-specific human video with only one ID image as reference.
ID-Animator: Zero-Shot Identity-Preserving Human Video Generation
Xuanhua He,
Quande Liu*,
Shengju Qian,
Xin Wang,
Tao Hu,
Ke Cao,
Keyu Yan,
Jie Zhang*
(*Corresponding Author)
Human Video Generation Demos
Recontextualization
| Reference Image | Output Video | Output Video | Output Video |
 |
 |
 |
 |
| Reference Image | Output Video | Output Video | Output Video |
 |
 |
 |
 |
Inference with Community Models
| Reference Image | Output Video | Output Video | Output Video |
 |
 |
 |
 |
| Reference Image | Output Video | Output Video | Output Video |
 |
 |
 |
 |
Identity Mixing
| Reference Image 1 | Reference Image 2 | Output Video | Output Video |
 |
 |
 |
 |
| Reference Image 1 | Reference Image 2 | Output Video | Output Video |
 |
 |
 |
 |
Combination with ControlNet
| Reference Image | Sketch Image | Output Video | Output Video |
|
 |
 |
 |
| Reference Image | Sketch Sequence | Output Video | Output Video |
|
 |
 |
 |
Contact Us
Xuanhua He: hexuanhua@mail.ustc.edu.cn
Quande Liu: qdliu0226@gmail.com
Shengju Qian: thesouthfrog@gmail.com
- Downloads last month
- 0
Unable to determine this model's library. Check the
docs
.