Research map / 3D vision / Human pose estimation
Encoder-Decoder With Multi-Level Attention for 3D Human Shape and Pose Estimation
Ziniu Wan, Zhengjia Li, Maoqing Tian, Jianbo Liu, Shuai Yi, Hongsheng Li
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-Temporal Encoder (parallel block) to the CNN with iterative decoder (ablation) | CNN + Iterative · 3DPW | PA-MPJPE: 52.2 -> 47.5 | Improved |
| add Kinematic Topology Decoder to the CNN encoder (ablation) | CNN + Iterative · 3DPW | PA-MPJPE: 52.2 -> 50.9 | Improved |
| add Kinematic Topology Decoder to the CNN encoder (ablation) | CNN + Iterative · 3DPW | MPJPE: 87.5 -> 88.0 | Hurt |
| add KTD on top of the STE encoder (ablation) | CNN+STE + Iterative · 3DPW | PA-MPJPE: 47.5 -> 45.7 | Improved |
| add KTD on top of the STE encoder (ablation) | CNN+STE + Iterative · 3DPW | MPJPE: 80.2 -> 79.1 | Improved |
| add temporal-only attention encoder (TE) (ablation) | CNN + Iterative · 3DPW | PA-MPJPE: 52.2 -> 51.1 | Improved |
| add spatial-only attention encoder (SE) (ablation) | CNN + Iterative · 3DPW | PA-MPJPE: 52.2 -> 49.8 | Improved |
| use parallel block with attentive addition instead of series block (ablation) | CNN+STE + Iterative · 3DPW | PA-MPJPE: 48.5 -> 47.5 | Improved |
| add attentive weights (v2) to parallel block instead of mean (v1) (ablation) | CNN+STE parallel + Iterative · 3DPW | PA-MPJPE: 48.1 -> 47.5 | Improved |
| use parallel block instead of coupling block (ablation) | CNN+STE + Iterative · 3DPW | PA-MPJPE: 49.3 -> 47.5 | Improved |
| add temporal modeling via coupling block over spatial-only encoder (ablation) | CNN+SE + Iterative · 3DPW | PA-MPJPE: 49.8 -> 49.3 | No clear effect |
| use vanilla transformer decoder instead of iterative regressor (ablation) | CNN+STE · 3DPW | PA-MPJPE: 47.5 -> 47.7 | No clear effect |
Earlier methods this paper compares with
Named baselines from earlier accepted papers in this paper's results tables: better than 2 of 2 on most shared results.
- VIBE — VIBE: Video Inference for Human Body Pose and Shape Estimation (CVPR 2020) · better on 10 of 10 shared results
- I2LMeshNet — I2L-MeshNet: Image-to-Lixel Prediction Network for Accurate 3D Human Pose and Mesh Estimation from a Single RGB Image (ECCV 2020) · better on 1 of 1 shared result
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
- SVTformer: Spatial-View-Temporal Transformer for Multi-View 3D Human Pose Estimation (AAAI 2025 · similarity 0.89)
- Attention Mechanism Exploits Temporal Contexts: Real-Time 3D Human Pose Reconstruction (CVPR 2020 · similarity 0.88)
- Global-to-Local Modeling for Video-Based 3D Human Pose and Shape Estimation (CVPR 2023 · similarity 0.88)
- Weakly-Supervised Discovery of Geometry-Aware Representation for 3D Human Pose Estimation (CVPR 2019 · similarity 0.88)
- Deep Semantic Graph Transformer for Multi-View 3D Human Pose Estimation (AAAI 2024 · similarity 0.86)
- HiPoser: 3D Human Pose Estimation with Hierarchical Shared Learning at Parts-Level Using Inertial Measurement Units (AAAI 2025 · similarity 0.86)
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.
