Research map / Video / Video generation
EMOTE: Expressive Motion and Shape Disentanglement for Human Animation
Dongbin Zhang, Hao Liu, Bingquan Dai, Kangjie Chen, Chuming Wang, Chen Li, Jing LYU, Haoqian Wang
Read the paperCheck your idea against 137,217 papers
AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.
Measured effects reported in the paper
| What was changed | Where | Measured | Effect |
|---|---|---|---|
| add proposed confidence-aware motion tracker (ablation) | EMOSH · EchoMimicV2 (self-driven) | PSNR: 20.51 -> 23.66 | Improved |
| add proposed confidence-aware motion tracker (ablation) | EMOSH · EchoMimicV2 (self-driven) | LPIPS: 0.1909 -> 0.1428 | Improved |
| add hybrid 2D keypoint injection on mesh condition (ablation) | EMOSH · EchoMimicV2 (self-driven) | PSNR: 22.90 -> 23.66 | Improved |
| add hybrid 2D keypoint injection (ablation) | EMOSH · Self-collected (self-driven) | LPIPS: 0.2973 -> 0.2933 | Improved |
| add motion-shape disentanglement via EHM retargeting (ablation) | EMOSH · Cross-driven (120 samples) | Identity Preservation Score: 0.4258 -> 0.4445 | Improved |
| add Spatially-Aligned Conditioning (ablation) | EMOSH · Cross-driven (120 samples) | Identity Preservation Score: 0.4237 -> 0.4445 | Improved |
| add motion-shape disentanglement (ablation) | EMOSH · Shape consistency (cross-driven) | Shape consistency score: 0.3968 -> 0.4483 | Improved |
| use frontal rather than side-view spatial anchor (ablation) | EMOSH · Cross-driven (120 samples) | IPS | No clear effect |
| use joint-optimization tracker instead of GUAVA cascaded tracker | motion tracker · Tracking 436-frame video | time (s): 424 -> 223 | Improved |
| use mesh (EHM) condition vs 2D pose (Wan-Animate) | EMOSH · TikTok (self-driven) | PSNR: 17.18 -> 17.80 | Improved |
| add rotational camera trajectories | EMOSH · Camera control | qualitative | Hurt |
Closest papers by meaning
- X-Dyna: Expressive Dynamic Human Image Animation (CVPR 2025 · similarity 0.89)
- HumanVid: Demystifying Training Data for Camera-controllable Human Image Animation (NeurIPS 2024 · similarity 0.87)
- Expressive Gaussian Human Avatars from Monocular RGB Video (NeurIPS 2024 · similarity 0.87)
- EMOPortraits: Emotion-enhanced Multimodal One-shot Head Avatars (CVPR 2024 · similarity 0.87)
- DisPose: Disentangling Pose Guidance for Controllable Human Image Animation (ICLR 2025 · similarity 0.87)
- VLOGGER: Multimodal Diffusion for Embodied Avatar Synthesis (CVPR 2025 · similarity 0.86)
Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.
