Research map / 3D vision / Human pose estimation
Egocentric Visibility-Aware Human Pose Estimation
Peng Dai, Yu Zhang, Feng Yiqiang, Zhen Fan, Yang Zhang
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 visibility-aware network and visibility-weighted loss (ablation) | EvaPose-ResNet50 · Eva-3M | MPJPE (mm): 40.6 -> 35.6 | Improved |
| add visibility-aware modeling (ablation) | EvaPose-ResNet50 · Eva-3M visible limb keypoints | VLK-PE (mm): 53.5 -> 45.8 | Improved |
| add visibility-aware modeling (ablation) | EvaPose-ResNet50 · Eva-3M invisible limb keypoints | ILK-PE (mm): 65.3 -> 65.6 | No clear effect |
| add VQ-VAE pose prior (ablation) | EvaPose-ResNet50 · Eva-3M | MPJPE (mm): 39.1 -> 35.6 | Improved |
| add VQ-VAE pose prior (ablation) | EvaPose-ResNet50 · Eva-3M | Jitter: 2.7 -> 3.0 | Hurt |
| add Stereo Transformer Decoder (ablation) | EvaPose-ResNet50 · Eva-3M | MPJPE (mm): 40.3 -> 39.1 | Improved |
| add Temporal Transformer Encoder (ablation) | EvaPose-ResNet50 · Eva-3M | MPJPE (mm): 46.0 -> 40.3 | Improved |
| add Temporal Transformer Encoder (ablation) | EvaPose-ResNet50 · Eva-3M | Jitter: 4.7 -> 2.5 | Improved |
| replace FRAME with EvaPose-ResNet50 | FRAME · EMHI P2 (unseen actions) | MPJPE (mm): 60.5 -> 38.5 | Improved |
| replace FRAME with EvaPose-ResNet50 | FRAME · EMHI P1 | PA-MPJPE (mm): 28.2 -> 28.0 | No clear effect |
| use ViT-L instead of ResNet50 backbone | EvaPose · Eva-3M | MPJPE (mm): 35.6 -> 34.2 | Improved |
Closest papers by meaning
- Scene-Aware Egocentric 3D Human Pose Estimation (CVPR 2023 · similarity 0.89)
- EMHI: A Multimodal Egocentric Human Motion Dataset with HMD and Body-Worn IMUs (AAAI 2025 · similarity 0.88)
- FRAME: Floor-aligned Representation for Avatar Motion from Egocentric Video (CVPR 2025 · similarity 0.87)
- 3D Human Pose Perception from Egocentric Stereo Videos (CVPR 2024 · similarity 0.86)
- EgoGen: An Egocentric Synthetic Data Generator (CVPR 2024 · similarity 0.86)
- EgoSim: An Egocentric Multi-view Simulator and Real Dataset for Body-worn Cameras during Motion and Activity (NeurIPS 2024 · similarity 0.85)
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.
