Research map / 3D vision / Human pose estimation
Mutual Learning to Adapt for Joint Human Parsing and Pose Estimation
Xuecheng Nie, Jiashi Feng, Shuicheng Yan
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 mutual learning-to-adapt module (ablation) | VGG16-FCN · LIP val (VGG16-FCN) | mIoU: 34.5 -> 40.2 | Improved |
| add mutual learning-to-adapt module (ablation) | VGG16-FCN · LIP val (VGG16-FCN) | PCK: 69.1 -> 76.0 | Improved |
| use adaptive parameters instead of direct addition fusion (ablation) | VGG16-FCN-MuLA · LIP val (VGG16-FCN) | mIoU: 36.5 -> 40.2 | Improved |
| use separate task networks with MuLA instead of tied multi-task learning (ablation) | VGG16-FCN · LIP val (VGG16-FCN) | PCK: 65.3 -> 76.0 | Improved |
| add dynamic convolution (mutual interaction) over plain residual conv (ablation) | VGG16-FCN-MuLA · LIP val (VGG16-FCN) | mIoU: 36.1 -> 40.2 | Improved |
| add pose-to-parsing adaptation in addition to parsing-to-pose (bidirectional) (ablation) | VGG16-FCN-LA-Pose · LIP val (VGG16-FCN) | mIoU: 32.1 -> 40.2 | Improved |
| add parsing-to-pose adaptation in addition to pose-to-parsing (bidirectional) (ablation) | VGG16-FCN-LA-Parsing · LIP val (VGG16-FCN) | PCK: 66.5 -> 76.0 | Improved |
| add mutual adaptation at the 1st stage of 2-stage Hourglass (ablation) | HG-2s-1u-MuLA · LIP val (Hourglass) | mIoU: 45.6 -> 46.9 | Improved |
| add mutual adaptation instead of plain stacking of Hourglass modules (ablation) | HG-2s-1u-MuLA · LIP val (Hourglass) | mIoU: 41.3 -> 46.9 | Improved |
| increase MuLA stages from 0 to 5 (ablation) | HG-MuLA · LIP val (Hourglass) | mIoU: 38.5 -> 49.3 | Improved |
| increase MuLA stages from 0 to 5 (ablation) | HG-MuLA · LIP val (Hourglass) | PCK: 78.8 -> 85.4 | Improved |
| add MuLA to Hourglass baseline | Hourglass · PASCAL-Person-Part parsing | mIoU: 62.9 -> 65.1 | Improved |
Closest papers by meaning
- Neural Architecture Search for Joint Human Parsing and Pose Estimation (ICCV 2021 · similarity 0.87)
- Human Pose Estimation With Parsing Induced Learner (CVPR 2018 · similarity 0.87)
- On the Correlation among Edge, Pose and Parsing (TPAMI 2021 · similarity 0.86)
- From Simple to Complex Scenes: Learning Robust Feature Representations for Accurate Human Parsing (TPAMI 2024 · similarity 0.84)
- Fast Adaptation for Human Pose Estimation via Meta-Optimization (CVPR 2024 · similarity 0.83)
- Referring Human Pose and Mask Estimation In the Wild (NeurIPS 2024 · 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.
