Research map / 3D vision / Human pose estimation
3D Human Pose Estimation With Spatial and Temporal Transformers
Ce Zheng, Sijie Zhu, Matias Mendieta, Taojiannan Yang, Chen Chen, Zhengming Ding
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 spatial transformer to temporal-only transformer (ablation) | PoseFormer · Human3.6M Protocol 1 (f=9) | MPJPE (mm): 52.5 to 49.9 | Improved |
| add temporal positional embedding to spatial-temporal model (ablation) | PoseFormer · Human3.6M Protocol 1 (f=9) | MPJPE (mm): 51.6 -> 50.5 | Improved |
| add spatial positional embedding to spatial-temporal model (ablation) | PoseFormer · Human3.6M Protocol 1 (f=9) | MPJPE (mm): 51.6 -> 50.7 | Improved |
| add both spatial and temporal positional embeddings (ablation) | PoseFormer · Human3.6M Protocol 1 (f=9) | MPJPE (mm): 51.6 -> 49.9 | Improved |
| add spatial transformer to temporal baseline (Protocol 1, f=81) (ablation) | PoseFormer · Human3.6M Protocol 1 | MPJPE (mm): 47.2 -> 44.3 | Improved |
| add spatial transformer to temporal baseline (Protocol 2, f=81) (ablation) | PoseFormer · Human3.6M Protocol 2 | P-MPJPE (mm): 37.0 -> 34.6 | Improved |
| increase input frames from 9 to 27 (ablation) | PoseFormer · Human3.6M Protocol 1 | MPJPE (mm): 49.9 -> 47.0 | Improved |
| increase input frames from 27 to 81 (ablation) | PoseFormer · Human3.6M Protocol 1 | MPJPE (mm): 47.0 -> 44.3 | Improved |
| use ground-truth 2D pose instead of CPN-detected 2D pose | PoseFormer · Human3.6M Protocol 1 | MPJPE (mm): 44.3 to 31.3 | Improved |
| increase layers L_S=L_T from 2 to 4 at c=32 (ablation) | PoseFormer · Human3.6M Protocol 1 | MPJPE (mm): 52.4 -> 49.9 | Improved |
| increase layers L_S=L_T from 4 to 6 at c=32 (ablation) | PoseFormer · Human3.6M Protocol 1 | MPJPE (mm): 49.9 -> 50.3 | Hurt |
| increase embedding dim c from 32 to 48 at 4 layers (ablation) | PoseFormer · Human3.6M Protocol 1 | MPJPE (mm): 49.9 -> 50.4 | Hurt |
Earlier methods this paper compares with
Named baselines from earlier accepted papers in this paper's results tables: better than 1 of 1 on most shared results.
- SRNet — SRNet: Improving Generalization in 3D Human Pose Estimation with a Split-and-Recombine Approach (ECCV 2020) · better on 2 of 2 shared results
Matched by method name in results tables read by AI. Authors' own tables tend to favour the newer method, and a comparison can be missed when a method is renamed.
Closest papers by meaning
- PoseFormerV2: Exploring Frequency Domain for Efficient and Robust 3D Human Pose Estimation (CVPR 2023 · similarity 0.92)
- A Single 2D Pose with Context is Worth Hundreds for 3D Human Pose Estimation (NeurIPS 2023 · similarity 0.91)
- SVTformer: Spatial-View-Temporal Transformer for Multi-View 3D Human Pose Estimation (AAAI 2025 · similarity 0.90)
- Hourglass Tokenizer for Efficient Transformer-Based 3D Human Pose Estimation (CVPR 2024 · similarity 0.89)
- Optimizing Human Pose Estimation Through Focused Human and Joint Regions (AAAI 2025 · similarity 0.88)
- Multiple View Geometry Transformers for 3D Human Pose Estimation (CVPR 2024 · similarity 0.88)
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.
