Saved in:
| Main Authors: | , , , , , , , , |
|---|---|
| Format: | Preprint |
| Published: |
2026
|
| Subjects: | |
| Online Access: | https://arxiv.org/abs/2602.07498 |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1866917256968011776 |
|---|---|
| author | Xu, Zhufeng Gao, Xuan Liu, Feng-Lin Zhang, Haoxian Fang, Zhixue Lai, Yu-Kun Liu, Xiaoqiang Wan, Pengfei Gao, Lin |
| author_facet | Xu, Zhufeng Gao, Xuan Liu, Feng-Lin Zhang, Haoxian Fang, Zhixue Lai, Yu-Kun Liu, Xiaoqiang Wan, Pengfei Gao, Lin |
| contents | Recent progress in video diffusion models has markedly advanced character animation, which synthesizes motioned videos by animating a static identity image according to a driving video. Explicit methods represent motion using skeleton, DWPose or other explicit structured signals, but struggle to handle spatial mismatches and varying body scales. %proportions. Implicit methods, on the other hand, capture high-level implicit motion semantics directly from the driving video, but suffer from identity leakage and entanglement between motion and appearance. To address the above challenges, we propose a novel implicit motion representation that compresses per-frame motion into compact 1D motion tokens. This design relaxes strict spatial constraints inherent in 2D representations and effectively prevents identity information leakage from the motion video. Furthermore, we design a temporally consistent mask token-based retargeting module that enforces a temporal training bottleneck, mitigating interference from the source images' motion and improving retargeting consistency. Our methodology employs a three-stage training strategy to enhance the training efficiency and ensure high fidelity. Extensive experiments demonstrate that our implicit motion representation and the propose IM-Animation's generative capabilities are achieve superior or competitive performance compared with state-of-the-art methods. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2602_07498 |
| institution | arXiv |
| publishDate | 2026 |
| record_format | arxiv |
| spellingShingle | IM-Animation: An Implicit Motion Representation for Identity-decoupled Character Animation Xu, Zhufeng Gao, Xuan Liu, Feng-Lin Zhang, Haoxian Fang, Zhixue Lai, Yu-Kun Liu, Xiaoqiang Wan, Pengfei Gao, Lin Computer Vision and Pattern Recognition Recent progress in video diffusion models has markedly advanced character animation, which synthesizes motioned videos by animating a static identity image according to a driving video. Explicit methods represent motion using skeleton, DWPose or other explicit structured signals, but struggle to handle spatial mismatches and varying body scales. %proportions. Implicit methods, on the other hand, capture high-level implicit motion semantics directly from the driving video, but suffer from identity leakage and entanglement between motion and appearance. To address the above challenges, we propose a novel implicit motion representation that compresses per-frame motion into compact 1D motion tokens. This design relaxes strict spatial constraints inherent in 2D representations and effectively prevents identity information leakage from the motion video. Furthermore, we design a temporally consistent mask token-based retargeting module that enforces a temporal training bottleneck, mitigating interference from the source images' motion and improving retargeting consistency. Our methodology employs a three-stage training strategy to enhance the training efficiency and ensure high fidelity. Extensive experiments demonstrate that our implicit motion representation and the propose IM-Animation's generative capabilities are achieve superior or competitive performance compared with state-of-the-art methods. |
| title | IM-Animation: An Implicit Motion Representation for Identity-decoupled Character Animation |
| topic | Computer Vision and Pattern Recognition |
| url | https://arxiv.org/abs/2602.07498 |