Research map / Video / Video generation
LieWarper: Geometry-Aware Motion Transfer via Lie Algebra
Linsong Shan, Laurence Tianruo Yang, Zecan Yang, Fukai Guo, Honglu Zhao, Yixuan Geng
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 |
|---|---|---|---|
| model motion on the Sim(2) manifold with analytic solver instead of baseline Euclidean motion guidance | LieWarper · DAVIS 2017 | FTD: 0.2497 -> 0.2381 vs Go-with-the-Flow | Improved |
| constrain to Sim(2) instead of general affine (6-DoF) (ablation) | LieWarper · DAVIS 2017 | Warp-Err: 24.1749 -> 19.3053 | Improved |
| add HOSVD filtering of input flow (ablation) | LieWarper · DAVIS 2017 | Depth-Corr: 0.8103 -> 0.8418 | Improved |
| add dense non-rigid flow residual to the rigid transform (ablation) | LieWarper · DAVIS 2017 | FTD: 0.3854 -> 0.2381 | Improved |
| add global rigid modulation to non-rigid flow (ablation) | LieWarper · DAVIS 2017 | FTD: 0.2501 -> 0.2381 | Improved |
| initialize with analytic solution rather than random init (ablation) | LieWarper · DAVIS 2017 | FTD: 0.4126 -> 0.3814 | Improved |
| add 5 fine-tuning steps after analytic init (ablation) | LieWarper · DAVIS 2017 | Warp-Err: 28.3225 -> 19.3053 | Improved |
| increase fine-tuning from 5 to 10 steps (ablation) | LieWarper · DAVIS 2017 | FTD: 0.2381 -> 0.2366; time 289.05 -> 533.25 s | No clear effect |
Earlier methods this paper compares with
Named baselines from earlier accepted papers in this paper's results tables: better than 2 of 2 on most shared results.
- RoPECraft — RoPECraft: Training-Free Motion Transfer with Trajectory-Guided RoPE Optimization on Diffusion Transformers (NeurIPS 2025) · better on 3 of 3 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
- FlowMotion: Training-Free Flow Guidance for Video Motion Transfer (CVPR 2026 · similarity 0.87)
- STANCE: Controllable Video Generation for Structured Dynamics via Sparse-To-dense ANChored Encoding (ECCV 2026 · similarity 0.87)
- Learning to Correct Geometry in Generated Videos (NeurIPS 2026 · similarity 0.86)
- Restage4D: Reanimating Deformable 3D Reconstruction from a Single Video (NeurIPS 2025 · similarity 0.86)
- HAIF-GS: Hierarchical and Induced Flow-Guided Gaussian Splatting for Dynamic Scene (NeurIPS 2025 · similarity 0.84)
- Wan-Move: Motion-controllable Video Generation via Latent Trajectory Guidance (NeurIPS 2025 · similarity 0.84)
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.
