Research map / Video / Video generation
EffiVMT: Video Motion Transfer via Efficient Spatial-Temporal Decoupled Finetuning
Yue Ma, Yulong Liu, Qiyuan Zhu, Xiangpeng Yang, Kunyu Feng, Xinhua Zhang, Zexuan Yan, Zhifeng Li, Sirui Han, Chenyang Qi, Qifeng Chen
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 spatial-temporal decoupled LoRA (ablation) | EffiVMT · MotionBench | Motion fidelity: 0.546 -> 0.971 | Improved |
| add adaptive RoPE (ablation) | EffiVMT · MotionBench | Motion fidelity: 0.655 -> 0.971 | Improved |
| add sparse motion sampling (ablation) | EffiVMT · MotionBench | tuning time (s): 2068 -> 727 | Improved |
| add sparse motion sampling (ablation) | EffiVMT · MotionBench | Motion fidelity: 0.975 -> 0.971 | No clear effect |
| add adaptive RoPE (ablation) | EffiVMT · MotionBench | Temporal consistency: 0.817 -> 0.976 | Improved |
| add spatial-temporal decoupled LoRA (ablation) | EffiVMT · MotionBench | tuning time (s): 971 -> 727 | Hurt |
| classify heads with pseudo-label attention maps instead of random assignment (ablation) | EffiVMT · MotionBench | Motion fidelity: 0.922 -> 0.971 | Improved |
| use alpha 1.25 instead of 0.75 for head classification (ablation) | EffiVMT · MotionBench | Motion fidelity: 0.925 -> 0.971 | Improved |
| use alpha 1.25 instead of 1.75 (ablation) | EffiVMT · MotionBench | Motion fidelity: 0.957 -> 0.971 | Improved |
| use EffiVMT instead of MotionDirector (WAN-2.1 re-implementation) | EffiVMT · MotionBench, Wan2.1 backbone | tuning time (s): 4641 -> 727 | Improved |
| switch backbone from CogVideoX to Wan2.1 | EffiVMT · MotionBench | Motion fidelity gain over MotionDirector: 0.016 -> 0.04 | Improved |
Earlier methods this paper compares with
Named baselines from earlier accepted papers in this paper's results tables: better than 1 of 1 on most shared results.
- MotionClone — MotionClone: Training-Free Motion Cloning for Controllable Video Generation (ICLR 2025) · better on 2 of 2 shared results
Matched by method name in results tables read by AI. Authors' own tables tend to favour the newer method, and a comparison can be missed when a method is renamed.
Closest papers by meaning
- Video Motion Transfer with Diffusion Transformers (CVPR 2025 · similarity 0.89)
- Extrapolating and Decoupling Image-to-Video Generation Models: Motion Modeling is Easier Than You Think (CVPR 2025 · similarity 0.87)
- EDEN: Enhanced Diffusion for High-quality Large-motion Video Frame Interpolation (CVPR 2025 · similarity 0.87)
- FlowMo: Variance-Based Flow Guidance for Coherent Motion in Video Generation (NeurIPS 2025 · similarity 0.87)
- Spectral Motion Alignment for Video Motion Transfer Using Diffusion Models (AAAI 2025 · similarity 0.87)
- VMC: Video Motion Customization using Temporal Attention Adaption for Text-to-Video Diffusion Models (CVPR 2024 · similarity 0.87)
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.
