Research map / Video / Video generation
MoCapAnything: Unified 3D Motion Capture for Arbitrary Skeletons from Monocular Videos
Kehong Gong, Zhengyu Wen, Weixia He, Mingxi Xu, Qi Wang, Ning Zhang, Zhengyu Li, Dongze Lian, Wei Zhao, Xiaoyu He, Mingyuan Zhang
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 the reference image-set encoder and cross-attention (ablation) | MoCapAnything · Truebones Zoo test (unseen species) | MPJPE (cm): 2.85 -> 1.76 | Improved |
| add mesh features to reference and video streams (ablation) | MoCapAnything · Truebones Zoo test (seen species) | MPJPE (cm): 1.88 -> 1.06 | Improved |
| add mesh features (ablation) | MoCapAnything · Truebones Zoo test (unseen species) | MPJPE (cm): 3.16 -> 1.76 | Improved |
| add mesh features (ablation) | MoCapAnything · Truebones Zoo test (unseen species) | MPJVE (cm): 0.44 -> 0.36 | Improved |
| add graph multi-head attention over skeleton topology (ablation) | MoCapAnything · Truebones Zoo test (rare species) | MPJPE (cm): 1.49 -> 1.28 | Improved |
| add graph multi-head attention over skeleton topology (ablation) | MoCapAnything · Truebones Zoo test (seen species) | MPJPE (cm): 1.08 -> 1.06 | No clear effect |
| add reference image set (ablation) | MoCapAnything · Truebones Zoo test (seen species) | MPJVE (cm): 0.68 -> 0.44 | Improved |
| increase encoder layers from 1 to 4 (ablation) | MoCapAnything · Truebones Zoo test (rare species) | MPJPE (cm): 1.90 -> 1.28 | Improved |
| increase encoder layers from 2 to 4 (ablation) | MoCapAnything · Truebones Zoo test (unseen species) | MPJPE (cm): 2.00 -> 1.76 | Improved |
| increase decoder layers from 12 to 16 (ablation) | MoCapAnything · Truebones Zoo test (seen species) | MPJPE (cm): 1.06 -> 1.05 | No clear effect |
| increase decoder layers from 12 to 16 (ablation) | MoCapAnything · Truebones Zoo test (rare species) | MPJPE (cm): 1.28 -> 1.46 | Hurt |
| use factorized joint-trajectory prediction plus IK versus adapted human pose and mocap baselines | MoCapAnything · Truebones Zoo test (unseen species) | MPJPE (cm): 7.42 (GLoT) -> 1.76 | Improved |
Closest papers by meaning
- MotionDreamer: Universal Skeletal Motion Generation for 3D Rigged Shapes (ECCV 2026 · similarity 0.89)
- Learning Monocular 3D Reconstruction of Articulated Categories From Motion (CVPR 2021 · similarity 0.88)
- SAMoR: Motion Modelling for Articulated Objects of Any Skeleton and Topology (NeurIPS 2026 · similarity 0.88)
- Learning Implicit Representation for Reconstructing Articulated Objects (ICLR 2024 · similarity 0.88)
- VideoMDM: Towards 3D Human Motion Generation From 2D Supervision (NeurIPS 2026 · similarity 0.87)
- S3O: A Dual-Phase Approach for Reconstructing Dynamic Shape and Skeleton of Articulated Objects from Single Monocular Video (ICML 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.
