Research map / 3D vision / Human pose estimation
Learning Knowledge from Textual Descriptions for 3D Human Pose Estimation
Yi Wu, Jingtian Li, Shangfei Wang, Guoming Li, Meng Mao, Linxiang Tan
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 pose text description alignment (cross-modality) | D3DP · Human3.6M | MPJPE (mm): 40.0 -> 39.4 | Improved |
| add pose text description alignment (cross-modality) | MixSTE · Human3.6M | MPJPE (mm): 40.9 -> 40.0 | Improved |
| add pose text description alignment (cross-modality) | D3DP · MPI-INF-3DHP | MPJPE (mm): 30.2 -> 29.3 | Improved |
| add FC alignment (ablation) | D3DP · Human3.6M | MPJPE (mm): 40.0 -> 39.7 | Improved |
| add transformer alignment instead of FC (ablation) | D3DP · Human3.6M | MPJPE (mm): 39.7 -> 39.6 | Improved |
| add cross-modality alignment instead of transformer (ablation) | D3DP · Human3.6M | MPJPE (mm): 39.6 -> 39.4 | Improved |
| add FC alignment (ablation) | MixSTE · Human3.6M | MPJPE (mm): 40.9 -> 40.6 | Improved |
| add cross-modality alignment instead of transformer (ablation) | MixSTE · Human3.6M | MPJPE (mm): 40.4 -> 40.0 | Improved |
| add transformer alignment instead of FC (ablation) | D3DP · MPI-INF-3DHP | MPJPE (mm): 29.7 -> 29.8 | Hurt |
| add cross-modality alignment instead of FC (ablation) | D3DP · MPI-INF-3DHP | MPJPE (mm): 29.7 -> 29.3 | Improved |
| add text-description supervision for edge-node regions (ablation) | D3DP · Human3.6M region R2 | MPJPE (mm) | Improved |
| add textual descriptions | D3DP · Human3.6M z-axis | absolute difference (z): 28.70 -> 27.92 | Improved |
Closest papers by meaning
- PoseScript: Linking 3D Human Poses and Natural Language (TPAMI 2024 · similarity 0.91)
- PoseScript: 3D Human Poses from Natural Language (ECCV 2022 · similarity 0.91)
- "PoseEmbroider: Towards a 3D, Visual, Semantic-aware Human Pose Representation" (ECCV 2024 · similarity 0.89)
- ChatPose: Chatting about 3D Human Pose (CVPR 2024 · similarity 0.88)
- A Simple yet Effective Baseline for 3D Human Pose Estimation (ICCV 2017 · similarity 0.87)
- Deep Semantic Graph Transformer for Multi-View 3D Human Pose Estimation (AAAI 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.
