Research map / 3D vision / Human pose estimation
MultiPoseNet: Fast Multi-Person Pose Estimation using Pose Residual Network
Muhammed Kocabas, Salih Karagoz, Emre Akbas
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 intermediate supervision (ablation) | R101 keypoint model · COCO val2017 | AP: 61.3 -> 63.9 | Improved |
| add concatenation of K features with final loss (ablation) | R101 keypoint model · COCO val2017 | AP: 62.1 -> 63.9 | Improved |
| use ResNet-101 instead of ResNet-50 (ablation) | MultiPoseNet · COCO val2017 | AP: 62.3 -> 63.9 | Improved |
| add dilated convolutions to R101 (ablation) | MultiPoseNet · COCO val2017 | AP: 63.9 -> 64.3 | Improved |
| add dropout to 1-layer 50 node PRN (ablation) | PRN · COCO val2017 (GT inputs) | AP: 76.3 -> 78.6 | Improved |
| increase PRN hidden nodes from 50 to 512 with dropout (ablation) | PRN · COCO val2017 (GT inputs) | AP: 78.6 -> 84.1 | Improved |
| use 1 layer instead of 2 layers of 512 nodes in PRN (ablation) | PRN · COCO val2017 (GT inputs) | AP: 81.9 -> 84.1 | Improved |
| use 1024 nodes with dropout and residual connection (ablation) | PRN · COCO val2017 (GT inputs) | AP: 84.1 -> 89.4 | Improved |
| use PRN instead of Max assignment (ablation) | MultiPoseNet · COCO val2017 | AP: 45.3 -> 64.3 | Improved |
| use PRN instead of UCR assignment (ablation) | MultiPoseNet · COCO val2017 | AP: 49.7 -> 64.3 | Improved |
| apply PRN to keypoints from Cao et al. | Cao et al. · COCO val2017 | AP: 58.4 -> 59.2 | Improved |
| apply PRN to keypoints from Newell et al. | Newell et al. · COCO val2017 | AP: 56.9 -> 58.1 | Improved |
Closest papers by meaning
- Simple Pose: Rethinking and Improving a Bottom-up Approach for Multi-Person Pose Estimation (AAAI 2020 · similarity 0.91)
- Cascaded Pyramid Network for Multi-Person Pose Estimation (CVPR 2018 · similarity 0.90)
- Towards Accurate Multi-Person Pose Estimation in the Wild (CVPR 2017 · similarity 0.90)
- Pose Partition Networks for Multi-Person Pose Estimation (ECCV 2018 · similarity 0.90)
- Dual Networks Based 3D Multi-Person Pose Estimation From Monocular Video (TPAMI 2022 · similarity 0.89)
- AdaptivePose: Human Parts as Adaptive Points (AAAI 2022 · 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.
