Research map / 3D vision / Human pose estimation
GraFormer: Graph-Oriented Transformer for 3D Pose Estimation
Weixi Zhao, Weiqiang Wang, Yunjie Tian
Check 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 |
|---|---|---|---|
| Use GraFormer instead of the reported baseline | 2D-to-3D pose estimation · Human3.6M | MPJPE (mm): 51.9 -> 51.8 | Improved |
| Use GraFormer instead of Zhao et al. | 2D-to-3D hand pose estimation · ObMan | MPJPE (mm): 2.34 -> 1.71 | Improved |
| Use GraFormer instead of the best shown baseline | 2D-to-3D hand pose estimation · FHAD | MPJPE (mm): 8.49 -> 10.42 | Hurt |
| Use GraFormer instead of Doosti et al. | 2D-to-3D hand pose estimation · GHD | MPJPE (mm): 8.45 -> 2.20 | Improved |
| Add GraAttention rather than removing it (ablation) | GraFormer · Human3.6M, ObMan, FHAD | MPJPE (mm): 47.81 -> 35.17; 8.51 -> 3.29; 16.30 -> 11.68 | Improved |
| Use GraAttention instead of self-attention (ablation) | GraFormer · Human3.6M, ObMan, FHAD | MPJPE (mm): 42.19 -> 35.17; 5.02 -> 3.29; 13.52 -> 11.68 | Improved |
| Add ChebGConv rather than removing it (ablation) | GraFormer · Human3.6M, ObMan, FHAD | MPJPE (mm): 37.78 -> 35.17; 7.29 -> 3.29; 14.39 -> 11.68 | Improved |
| Use graph convolution instead of retaining the MLP (ablation) | GraFormer · Human3.6M, ObMan, FHAD | MPJPE (mm): 42.44 -> 35.17; 3.46 -> 3.29; 13.49 -> 11.68 | Improved |
| Set ChebGConv order to 2 instead of 1 or 3 (ablation) | GraFormer · Human3.6M | MPJPE (mm): 39.04 -> 36.19; 39.31 -> 36.19 | Improved |
Closest papers by meaning
- Mesh Graphormer (ICCV 2021 · similarity 0.90)
- Multiple View Geometry Transformers for 3D Human Pose Estimation (CVPR 2024 · similarity 0.86)
- Deep Semantic Graph Transformer for Multi-View 3D Human Pose Estimation (AAAI 2024 · similarity 0.86)
- SVTformer: Spatial-View-Temporal Transformer for Multi-View 3D Human Pose Estimation (AAAI 2025 · similarity 0.86)
- GRPose: Learning Graph Relations for Human Image Generation with Pose Priors (AAAI 2025 · similarity 0.85)
- Learning to Filter Outlier Edges in Global SfM (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.
