Research map / Detection and segmentation / Weakly supervised detection and segmentation
Hierarchical Scene Parsing by Weakly Supervised Learning with Image Descriptions
Ruimao Zhang, Liang Lin, Guangrun Wang, Meng Wang, Wangmeng Zuo
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 multi-scale information (ablation) | CNN-RsNN · VOC 2012 val (weak) | mean IoU: 34.3 -> 34.7 | Improved |
| add contextual information (ablation) | CNN-RsNN · VOC 2012 val (weak) | mean IoU: 34.3 -> 34.4 | No clear effect |
| add both context and multi-scale (Full) (ablation) | CNN-RsNN · VOC 2012 val (weak) | mean IoU: 34.3 -> 35.1 | Improved |
| add contextual information (ablation) | CNN-RsNN · SYSU-Scenes (weak) | mean IoU: 21.5 -> 22.5 | Improved |
| add context and multi-scale (Full) (ablation) | CNN-RsNN · SYSU-Scenes (weak) | mean IoU: 21.5 -> 23.7 | Improved |
| use sentence-derived structure (Full) instead of DeepLab | DeepLab-EM-Adapt · VOC 2012 val (weak) | mean IoU: 30.3 -> 35.1 | Improved |
| add 280 strongly annotated images | CNN-RsNN Full · VOC 2012 val | mean IoU: 35.1 -> 43.7 | Improved |
| add 1464 strongly annotated images | CNN-RsNN Full · VOC 2012 val | mean IoU: 35.1 -> 51.7 | Improved |
| add 500 strongly annotated images | CNN-RsNN Full · SYSU-Scenes | mean IoU: 23.7 -> 29.7 | Improved |
| add 1241 strongly annotated images | CNN-RsNN Full · SYSU-Scenes | mean IoU: 23.7 -> 35.5 | Improved |
| add contextual information (joint CNN+RsNN update) (ablation) | CNN-RsNN · VOC 2012 val | structure accuracy: 64.2 -> 67.4 | Improved |
| update CNN jointly with RsNN instead of fixing CNN (with context) (ablation) | CNN-RsNN · VOC 2012 val | structure accuracy: 62.8 -> 67.4 | Improved |
Closest papers by meaning
- Deep Structured Scene Parsing by Learning With Image Descriptions (CVPR 2016 · similarity 0.97)
- Weakly Supervised Scene Parsing with Point-Based Distance Metric Learning (AAAI 2019 · similarity 0.90)
- Weakly Supervised Visual Semantic Parsing (CVPR 2020 · similarity 0.90)
- Combining Bottom-Up, Top-Down, and Smoothness Cues for Weakly Supervised Image Segmentation (CVPR 2017 · similarity 0.89)
- Learning To Generate Scene Graph From Natural Language Supervision (ICCV 2021 · similarity 0.89)
- Boosting Weakly Supervised Referring Image Segmentation via Progressive Comprehension (NeurIPS 2024 · 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.
