Research map / 3D vision / Human pose estimation
Exploiting Spatial-Temporal Relationships for 3D Pose Estimation via Graph Convolutional Networks
Yujun Cai, Liuhao Ge, Jun Liu, Jianfei Cai, Tat-Jen Cham, Junsong Yuan, Nadia Magnenat Thalmann
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 |
|---|---|---|---|
| split neighbours into temporal forward, backward and other classes with individual kernel weights (ablation) | uniform GCN · Human3.6M (3 frames) | MPJPE (mm): 69.8 -> 54.8 | Improved |
| add symmetrical node class (ablation) | split temporal GCN · Human3.6M (3 frames) | MPJPE (mm): 54.8 -> 54.0 | Improved |
| add physical connection class (ablation) | split temporal & symmetrical GCN · Human3.6M (3 frames) | MPJPE (mm): 54.0 -> 49.1 | Improved |
| add last pooling and first upsampling layers (ablation) | local-to-global network · Human3.6M Protocol #1 | MPJPE (mm): 52.3 -> 49.1 | Improved |
| add all pooling and upsampling layers (ablation) | local-to-global network · Human3.6M Protocol #1 | MPJPE (mm): 53.9 -> 49.1 | Improved |
| increase input frames from 1 to 3 (ablation) | graph convolution model · Human3.6M | MPJPE (mm): 50.62 -> 49.08 | Improved |
| increase input frames from 3 to 7 (ablation) | graph convolution model · Human3.6M | MPJPE (mm): 49.08 -> 48.78 | No clear effect |
| increase input frames from 1 to 7 (ablation) | graph convolution model · STB | MPJPE (mm): 6.95 -> 6.61 | Improved |
| add pose refinement (ablation) | graph convolution model · Human3.6M | MPJPE | Improved |
Closest papers by meaning
- HOPE-Net: A Graph-Based Model for Hand-Object Pose Estimation (CVPR 2020 · similarity 0.91)
- 3D Hand Shape and Pose Estimation From a Single RGB Image (CVPR 2019 · similarity 0.91)
- Optimizing Network Structure for 3D Human Pose Estimation (ICCV 2019 · similarity 0.90)
- Feature Boosting Network For 3D Pose Estimation (TPAMI 2019 · similarity 0.90)
- Learning to Estimate 3D Hand Pose From Single RGB Images (ICCV 2017 · similarity 0.90)
- Deep Semantic Graph Transformer for Multi-View 3D Human Pose Estimation (AAAI 2024 · similarity 0.89)
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.
