Research map / Detection and segmentation / CNN features
Learning Compositional Neural Information Fusion for Human Parsing
Wenguan Wang, Zhijie Zhang, Siyuan Qi, Jianbing Shen, Yanwei Pang, Ling Shao
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 |
|---|---|---|---|
| parse the whole human hierarchy with direct inference instead of only leaf parts (ablation) | backbone direct inference · PASCAL-Person-Part test | mIoU (1st level): 64.14 -> 65.27 | Improved |
| add bottom-up inference to direct inference (ablation) | direct model · PASCAL-Person-Part test | mIoU (1st level): 65.27 -> 65.42 | Improved |
| add top-down inference to direct inference (ablation) | direct model · PASCAL-Person-Part test | mIoU (1st level): 65.27 -> 69.02 | Improved |
| add bottom-up inference on top of direct + top-down (ablation) | direct + top-down · PASCAL-Person-Part test | mIoU (1st level): 69.02 -> 69.43 | Improved |
| add bottom-up inference on direct + top-down for 3rd level full body (ablation) | direct + top-down · PASCAL-Person-Part test (3rd level) | mIoU: 88.40 -> 91.02 | Improved |
| add confidence-conditioned fusion gates (ablation) | direct + bottom-up + top-down · PASCAL-Person-Part test | mIoU (1st level): 69.43 -> 70.76 | Improved |
| add confidence-conditioned fusion gates (ablation) | direct + bottom-up + top-down · PASCAL-Person-Part test (2nd level) | mIoU: 80.34 -> 81.62 | Improved |
| add confidence-conditioned fusion gates (ablation) | direct + bottom-up + top-down · PASCAL-Person-Part test (3rd level) | mIoU: 91.02 -> 91.31 | Improved |
| use full compositional fusion model instead of CE2P | human parser · LIP val | mIoU: 53.10 -> 57.74 | Improved |
| use full compositional fusion model instead of MuLA | human parser · LIP val | pixel accuracy: 88.5 -> 88.03 | Hurt |
| use full compositional fusion model instead of TGPNet | human parser · Fashion Clothing test | F-1 score: 51.92 -> 58.12 | Improved |
Closest papers by meaning
- Hierarchical Human Semantic Parsing with Comprehensive Part-Relation Modeling (TPAMI 2021 · similarity 0.93)
- CDGNet: Class Distribution Guided Network for Human Parsing (CVPR 2022 · similarity 0.89)
- Hierarchical Human Parsing With Typed Part-Relation Reasoning (CVPR 2020 · similarity 0.88)
- Learning Semantic Neural Tree for Human Parsing (ECCV 2020 · similarity 0.88)
- Part-Aware Context Network for Human Parsing (CVPR 2020 · similarity 0.87)
- Blended Grammar Network for Human Parsing (ECCV 2020 · similarity 0.87)
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.
