Research map / Image generation / GANs
Macro-Micro Adversarial Network for Human Parsing
Yawei Luo, Zhedong Zheng, Liang Zheng, Tao Guan, Junqing Yu, Yi Yang
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 Macro D to the baseline (ablation) | Do-DeepLab-ASPP · LIP val | mIoU: +0.88% mIoU | Improved |
| add Micro D to the baseline (ablation) | Do-DeepLab-ASPP · LIP val | mIoU: +0.80% mIoU | Improved |
| add Micro D on top of Macro D (full MMAN) (ablation) | MMAN · LIP val | mIoU: drops 1.21% when Macro D removed... 45.60 -> 46.81 | Improved |
| add Macro D on top of Micro D (full MMAN) (ablation) | MMAN · LIP val | mIoU: 45.52 -> 46.81 | Improved |
| add both discriminators to the baseline on PASCAL-Person-Part (ablation) | Do-DeepLab-ASPP · PASCAL-Person-Part test | mIoU: 56.79 -> 58.45 | Improved |
| use Macro D plus Micro D instead of a single discriminator (ablation) | MMAN · LIP val | mIoU: 45.23 -> 46.81 | Improved |
| place Macro D on deep layers instead of the top layer (vs Double AN) (ablation) | MMAN · LIP val | mIoU: +0.82% mIoU | Improved |
| add a third discriminator on the 3rd deconvolution layer (ablation) | MMAN · LIP val | mIoU: +0.16% mIoU | No clear effect |
| add the multi-scale attention network to MMAN | MMAN · PASCAL-Person-Part test | mIoU: 58.45 -> 59.91 | Improved |
| use MMAN instead of ASN adversarial loss on same baseline | Do-DeepLab-ASPP · LIP val | mIoU: +1.40% | Improved |
| use MMAN instead of SSL on same baseline | Do-DeepLab-ASPP · LIP val | mIoU: +0.62% | Improved |
| use MMAN over ASN for cross-dataset transfer | MMAN · PPSS (LIP-trained, merged labels) | mIoU: 50.7 -> 52.1 | Improved |
Closest papers by meaning
- Cross-Domain Human Parsing via Adversarial Feature and Label Adaptation (AAAI 2018 · similarity 0.89)
- From Simple to Complex Scenes: Learning Robust Feature Representations for Accurate Human Parsing (TPAMI 2024 · similarity 0.88)
- Deep Learning Technique for Human Parsing: A Survey and Outlook (IJCV 2024 · similarity 0.88)
- End-to-End One-Shot Human Parsing (TPAMI 2023 · similarity 0.87)
- Human Parsing with Contextualized Convolutional Neural Network (TPAMI 2016 · similarity 0.87)
- Progressive Cognitive Human Parsing (AAAI 2018 · 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.
