Research map / Detection and segmentation / Weakly supervised detection and segmentation
Pseudo-Mask Matters in Weakly-Supervised Semantic Segmentation
Yi Li, Zhanghui Kuang, Liyang Liu, Yimin Chen, Wayne Zhang
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-crop at train and test for CAM (ablation) | SEAM CAM, ResNet38 · PASCAL VOC 2012 val (CAM mIoU) | mIoU: 52.72 -> 56.21 | Improved |
| swap ResNet38 for ScaleNet101 in CAM classifier (multi-crop test) (ablation) | CAM · PASCAL VOC 2012 val (CAM mIoU) | mIoU: 53.93 -> 57.81 | Improved |
| add dense-CRF to CAM (ablation) | our CAM · PASCAL VOC 2012 val (CAM mIoU) | mIoU: 58.21 -> 59.19 | Improved |
| add Coefficient of Variation Smoothing to CRF (ablation) | our CAM · PASCAL VOC 2012 val (CAM mIoU) | mIoU: 59.19 -> 60.83 | Improved |
| add CVS to PPMG (ablation) | our CAM · PASCAL VOC 2012 val (CAM mIoU) | mIoU: 61.23 -> 61.49 | Improved |
| use PPMG instead of CRF argmax (ablation) | our CAM · PASCAL VOC 2012 val (CAM mIoU) | mIoU: 59.19 -> 61.49 | Improved |
| add affinity network and Random Walk (ablation) | our CAM · PASCAL VOC 2012 val (CAM mIoU) | mIoU: 58.21 -> 52.18 | Hurt |
| add Random Walk to PPMG (ablation) | our CAM · PASCAL VOC 2012 val (CAM mIoU) | mIoU: 61.49 -> 60.96 | Hurt |
| add Pretended Under-fitting (clamp) to segmentation training (ablation) | PMM · PASCAL VOC 2012 val | mIoU: 64.86 -> 66.73 | Improved |
| use clamp instead of pow in PUS (ablation) | PMM · PASCAL VOC 2012 val | mIoU: 64.92 -> 66.73 | Improved |
| apply PUS clamp to SEAM with Random Walk pseudo-mask (ablation) | SEAM pseudo-mask · PASCAL VOC 2012 val | mIoU: 64.41 -> 63.27 | Hurt |
| add one round of cyclic pseudo-mask update (ablation) | PMM · PASCAL VOC 2012 val | mIoU: 66.73 -> 68.50 | Improved |
Closest papers by meaning
- Know Your Attention Maps: Class-specific Token Masking for Weakly Supervised Semantic Segmentation (ICCV 2025 · similarity 0.92)
- Weakly Supervised Semantic Segmentation via Progressive Confidence Region Expansion (CVPR 2025 · similarity 0.90)
- W2P: Switching from Weak Supervision to Partial Supervision for Semantic Segmentation (AAAI 2024 · similarity 0.90)
- From SAM to CAMs: Exploring Segment Anything Model for Weakly Supervised Semantic Segmentation (CVPR 2024 · similarity 0.90)
- Progressive Feature Self-Reinforcement for Weakly Supervised Semantic Segmentation (AAAI 2024 · similarity 0.90)
- Weakly Supervised Semantic Segmentation for Driving Scenes (AAAI 2024 · similarity 0.88)
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.
