Research map / Video / Video generation
Temporal Residual Jacobians for Rig-free Motion Transfer
Sanjeev Muralikrishnan, Niladri Shekhar Dutt, Siddhartha Chaudhuri, Noam Aigerman, Vladimir Kim, Matthew Fisher, Niloy Mitra
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 |
|---|---|---|---|
| predict temporal residual Jacobians with a neural ODE on top of per-frame NJF posing | NJF(M_t) · AMASS Walk | L2-V (cm): 6.85 -> 1.48 | Improved |
| predict temporal residual Jacobians on top of NJF | NJF(M_t) · AMASS Jump | L2-V (cm): 5.52 -> 2.64 | Improved |
| predict Jacobians (triangle deformations) instead of vertex displacements in the neural ODE | VertexODE · AMASS Run | L2-V (cm): 14.23 -> 1.73 | Improved |
| predict residual deformations over a base posing model instead of making a neural ODE predict full per-frame deformations (ablation) | Temporal Residual Jacobians · AMASS | motion transfer quality | Improved |
| condition posing network on first-frame Jacobians with a residual connection (ablation) | posing network · AMASS | pose prediction and generalization | Improved |
| use higher-order ODE solvers (Runge-Kutta, Runge-Kutta-Fehlberg) instead of Euler (ablation) | Temporal Residual Jacobians · AMASS | motion quality | No clear effect |
| express residual Jacobians relative to previous predicted frame instead of base Jacobians (ablation) | Temporal Residual Jacobians · AMASS | training feasibility | Hurt |
| use neural ODE alone to predict the whole sequence without a base posing network (ablation) | neural ODE · long motion sequences | drift | Hurt |
Closest papers by meaning
- JacobianAvatar: Temporally Consistent Semi-rigid Avatar Reconstruction from a Monocular Video (ECCV 2026 · similarity 0.87)
- LART: Neural Correspondence Learning with Latent Regularization Transformer for 3D Motion Transfer (NeurIPS 2023 · similarity 0.87)
- Mesh4D: 4D Mesh Reconstruction and Tracking from Monocular Video (CVPR 2026 · similarity 0.87)
- High-Fidelity Neural Human Motion Transfer From Monocular Video (CVPR 2021 · similarity 0.87)
- Motion2VecSets: Non-Rigid Shape Reconstruction and Tracking With 4D Latent Set Diffusion (TPAMI 2026 · similarity 0.86)
- Motion2VecSets: 4D Latent Vector Set Diffusion for Non-rigid Shape Reconstruction and Tracking (CVPR 2024 · similarity 0.85)
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.
