Research map / 3D vision / Human pose estimation
3D Human Pose Estimation Using Spatio-Temporal Networks with Explicit Occlusion Training
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 |
|---|---|---|---|
| add heat map embedding (ablation) | 3D pose TCN · Human3.6M | MPJPE (Protocol #1): 51.7 -> 51.0 | Helped |
| add multi-stride TCN (ablation) | 3D pose TCN · Human3.6M | MPJPE (Protocol #1): 51.0 -> 48.6 | Helped |
| add multi-view loss (ablation) | 3D pose TCN · Human3.6M | MPJPE (Protocol #1): 48.6 -> 47.3 | Helped |
| add spatial KCS (ablation) | 3D pose TCN · Human3.6M | MPJPE (Protocol #1): 47.3 -> 44.9 | Helped |
| add temporal KCS (ablation) | 3D pose TCN with spatial KCS · Human3.6M | MPJPE (Protocol #1): 44.9 -> 41.2 | Helped |
| add semi-supervised 2D data (ablation) | 3D pose TCN with KCS · Human3.6M | MPJPE (Protocol #1): 41.2 -> 40.1 | Helped |
| increase temporal strides from 1 to 1,2,3,5,7 (ablation) | 3D pose TCN · Human3.6M | MPJPE (Protocol #1): 45.4 -> 41.2 | Helped |
| increase temporal length from 8 to 64 (ablation) | 3D pose TCN · Human3.6M | MPJPE (Protocol #1): 50.2 -> 42.6 | Helped |
| increase embedding dimension from 256 to 512 (ablation) | heat map embedding · Human3.6M | MPJPE (Protocol #1): 43.1 -> 42.6 | Helped |
| increase embedding dimension from 512 to 1024 (ablation) | heat map embedding · Human3.6M | MPJPE (Protocol #1): 42.6 -> 42.9 | No clear effect |
| increase temporal interval for TKCS from 1 to 3 (ablation) | spatio-temporal discriminator · Human3.6M | MPJPE (Protocol #1): 42.6 -> 43.1 | Hurt |
| increase temporal interval for TKCS from 1 to 5 (ablation) | spatio-temporal discriminator · Human3.6M | MPJPE (Protocol #1): 42.6 -> 44.0 | Hurt |
Closest papers by meaning
- Capturing Humans in Motion: Temporal-Attentive 3D Human Pose and Shape Estimation From Monocular Video (CVPR 2022 · similarity 0.92)
- Graph and Temporal Convolutional Networks for 3D Multi-person Pose Estimation in Monocular Videos (AAAI 2021 · similarity 0.92)
- Deep Dual Consecutive Network for Human Pose Estimation (CVPR 2021 · similarity 0.91)
- 3D Human Pose Estimation in Video With Temporal Convolutions and Semi-Supervised Training (CVPR 2019 · similarity 0.90)
- Video-Based Human Pose Regression via Decoupled Space-Time Aggregation (CVPR 2024 · similarity 0.89)
- Optimizing Human Pose Estimation Through Focused Human and Joint Regions (AAAI 2025 · 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.