Research map / 3D vision / Human pose estimation
3D Human Pose Estimation With Spatio-Temporal Criss-Cross Attention
Read the paperCheck 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 |
|---|---|---|---|
| Use parallel spatial and temporal attention rather than spatial attention alone (ablation) | STCFormer · Human3.6M | P1 (mm): 275.5 -> 57.0 | Helped |
| Use parallel spatial and temporal attention rather than temporal attention alone (ablation) | STCFormer · Human3.6M | P1 (mm): 67.6 -> 57.0 | Helped |
| Add part-aware positional embedding (SPE1) (ablation) | STCFormer · Human3.6M | P1 (mm): 57.0 -> 56.4 | Helped |
| Add spatio-temporal convolutional positional embedding (SPE2) (ablation) | STCFormer · Human3.6M | P1 (mm): 57.0 -> 44.9 | Helped |
| Add both SPE1 and SPE2 (ablation) | STCFormer · Human3.6M | P1 (mm): 57.0 -> 44.1 | Helped |
| Use part-aware positional embedding instead of no SPE1 (ablation) | STCFormer · Human3.6M | P1 (mm): 48.7 -> 48.3 | Helped |
| Replace SPE1 with absolute positional embedding (ablation) | STCFormer · Human3.6M | P1 (mm): 48.3 -> 48.9 | Hurt |
| Replace SPE1 with centrality positional embedding (ablation) | STCFormer · Human3.6M | P1 (mm): 48.3 -> 49.6 | Hurt |
| Replace SPE1 with symmetric positional embedding (ablation) | STCFormer · Human3.6M | P1 (mm): 48.3 -> 49.2 | Hurt |
| Increase the input sequence from 27 to 243 frames (ablation) | STCFormer · Human3.6M | P1 (mm): 44.1 -> 41.0 | Helped |
| Increase hidden embedding channel size from 256 to 512 (ablation) | STCFormer · Human3.6M | P1 (mm): 41.0 -> 40.5 | Helped |
| Use STCFormer-L instead of MixSTE with 243 input frames | 3D pose estimation · Human3.6M (CPN 2D poses) | P1 (mm): 40.9 -> 40.5 | Helped |
Closest papers by meaning
- Optimizing Human Pose Estimation Through Focused Human and Joint Regions (AAAI 2025 · similarity 0.86)
- SVTformer: Spatial-View-Temporal Transformer for Multi-View 3D Human Pose Estimation (AAAI 2025 · similarity 0.84)
- DynPose: Largely Improving the Efficiency of Human Pose Estimation by a Simple Dynamic Framework (CVPR 2025 · similarity 0.84)
- MV-SSM: Multi-View State Space Modeling for 3D Human Pose Estimation (CVPR 2025 · similarity 0.84)
- FreeMan: Towards Benchmarking 3D Human Pose Estimation under Real-World Conditions (CVPR 2024 · similarity 0.83)
- PoseMamba: Monocular 3D Human Pose Estimation with Bidirectional Global-Local Spatio-Temporal State Space Model (AAAI 2025 · similarity 0.83)
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.