Research map / 3D vision / Human pose estimation
LCollision: Fast Generation of Collision-Free Human Poses using Learned Non-Penetration Constraints
Qingyang Tan, Zherong Pan, Dinesh Manocha
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 |
|---|---|---|---|
| decompose collisions into body-part domains (ablation) | LCollision · SCAPE, M=5×10^4 | CLASSIFY: 80.2% -> 82.8% | Improved |
| add penetration-depth loss (ablation) | LCollision · SCAPE, M=5×10^4 | CLASSIFY: 80.7% -> 82.8% | Improved |
| add ranking loss (ablation) | LCollision · SCAPE, M=5×10^4 | CLASSIFY: 81.1% -> 82.8% | Improved |
| add ranking loss (ablation) | LCollision · SCAPE, M=5×10^4 | MSE: 5.06×10^{-4} -> 6.72×10^{-4} | Hurt |
| increase augmented training set from M=5×10^4 to M=2.5×10^6 | LCollision · SCAPE | CLASSIFY: 82.8% -> 94.1% | Improved |
| increase augmented training set from M=5×10^4 to M=2.5×10^6 | LCollision · Swing | CLASSIFY: 91.2% -> 96.2% | Improved |
| increase augmented training set from M=5×10^4 to M=2.5×10^6 | LCollision · Jumping | CLASSIFY: 91.6% -> 95.4% | Improved |
| use learned collision detection instead of exact collision checking | LCollision CPU · SCAPE collision detection | speedup: 21x | Improved |
| use learned collision detection instead of exact collision checking | LCollision GPU · Swing collision detection | speedup: 342x | Improved |
Closest papers by meaning
- HuCollisionField: Resolving Self-Collisions via Neural Fields for Human Prediction (ECCV 2026 · similarity 0.89)
- N-Penetrate: Active Learning of Neural Collision Handler for Complex 3D Mesh Deformations (ICML 2022 · similarity 0.85)
- LCR-Net++: Multi-person 2D and 3D Pose Detection in Natural Images (TPAMI 2019 · similarity 0.85)
- CLOAF: CoLlisiOn-Aware Human Flow (CVPR 2024 · similarity 0.85)
- ManiPose: Manifold-Constrained Multi-Hypothesis 3D Human Pose Estimation (NeurIPS 2024 · similarity 0.84)
- FinePOSE: Fine-Grained Prompt-Driven 3D Human Pose Estimation via Diffusion Models (CVPR 2024 · similarity 0.83)
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.
