Research map / 3D vision / Face and shape
LEAP: Learning Articulated Occupancy of People
Check your idea against 108,967 papers
AI-extracted summary. Written by AI from the paper's title only (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 |
|---|---|---|---|
| replace per-subject occupancy modeling with canonicalized occupancy and learned LBS | LEAP versus NASA · DFaust unseen poses | IOU: 73.69 -> 98.39 | Helped |
| replace per-subject occupancy modeling with canonicalized occupancy and learned LBS | LEAP versus NASA · DFaust unseen poses | Chamfer distance: 4.72 -> 2.27 | Helped |
| replace per-subject occupancy modeling with canonicalized occupancy and learned LBS | LEAP versus NASA · MoVi unseen subjects | IOU: 78.87 -> 92.97 | Helped |
| replace per-subject occupancy modeling with canonicalized occupancy and learned LBS | LEAP versus NASA · MoVi unseen subjects | Chamfer distance: 3.67 -> 2.80 | Helped |
| optimize body placement with LEAP point-based collision loss | PLACE-generated bodies · Replica Room 1 | scene-human collision score: 3.51% -> 0.62% | Helped |
| optimize body placement with LEAP point-based collision loss | PLACE-generated bodies · Replica Room 1 | human-human collision score: 5.73% -> 1.06% | Helped |
| optimize body placement with LEAP point-based collision loss | PLACE-generated bodies · Replica Room 1 | human-scene collision score: 5.72% -> 5.72% | No clear effect |
| add Structure encoder (ablation) | LEAP feature encoders · DFaust unseen sequences | IOU: Shape + Structure: 97.96%; Shape: 96.44% | Helped |
| add Shape encoder (ablation) | LEAP feature encoders · DFaust unseen sequences | IOU: Shape + Structure: 97.96%; Structure: 97.49% | Helped |
| add Pose encoder to Shape + Structure (ablation) | LEAP feature encoders · DFaust unseen sequences | IOU: 97.96% -> 97.99% | Helped |
| use all three feature encoders instead of Pose encoder only (ablation) | LEAP feature encoders · DFaust unseen sequences | IOU: 91.86% -> 97.99% | Helped |
| add cycle-distance feature (ablation) | LEAP occupancy network · DFaust multi-person occupancy | reconstruction quality | Helped |
Closest papers by meaning
- COAP: Compositional Articulated Occupancy of People (CVPR 2022 · similarity 0.91)
- $\text{Di}^2\text{Pose}$: Discrete Diffusion Model for Occluded 3D Human Pose Estimation (NeurIPS 2024 · similarity 0.86)
- LEAP: Liberate Sparse-View 3D Modeling from Camera Poses (ICLR 2024 · similarity 0.86)
- Learning Implicit Representation for Reconstructing Articulated Objects (ICLR 2024 · similarity 0.85)
- Neural Localizer Fields for Continuous 3D Human Pose and Shape Estimation (NeurIPS 2024 · similarity 0.85)
- HiPART: Hierarchical Pose AutoRegressive Transformer for Occluded 3D Human Pose Estimation (CVPR 2025 · 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.