Research map / Detection and segmentation / Weakly supervised detection and segmentation
Weakly Supervised Instance Segmentation Using Class Peak Response
Read the paperCheck your idea against 108,967 papers
AI-extracted summary. Written by AI from the paper's title only (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 peak stimulation (ablation) | PRM · PASCAL VOC 2012 val. (pointwise localization) | mAP%: 81.5 -> 85.5 | Helped |
| add peak stimulation (ablation) | PRM · MS COCO 2014 val. (pointwise localization) | mAP%: 53.1 -> 57.5 | Helped |
| use peak stimulation instead of global max-min pooling | ResNet50 response aggregation · VOC 2012 segmentation val. (PRM quality) | mAP: 62.4 -> 64.0 | Helped |
| add PRM-based instance mask extraction | weakly supervised semantic segmentation · VOC 2012 val. | mIoU: 52.8 -> 53.4 | Helped |
| use PRM instead of MELM with MCG proposals | weakly supervised instance segmentation · PASCAL VOC 2012 val. | mAP^r_0.5: 22.9 -> 26.8 | Helped |
| add peak stimulation (ablation) | PRM instance segmentation with ResNet50 · PASCAL VOC 2012 val. | mAP^r_0.5: 22.8 -> 26.8 | Helped |
| add peak stimulation (ablation) | PRM instance segmentation with VGG16 · PASCAL VOC 2012 val. | mAP^r_0.5: 11.9 -> 22.0 | Helped |
| add instance-aware term (ablation) | PRM proposal retrieval · PASCAL VOC 2012 val. | mAP^r_0.5: 13.3 -> 26.8 | Helped |
| add class-aware term (ablation) | PRM proposal retrieval · PASCAL VOC 2012 val. | mAP^r_0.5: 16.5 -> 26.8 | Helped |
| add boundary-aware term (ablation) | PRM proposal retrieval · PASCAL VOC 2012 val. | mAP^r_0.5: 24.3 -> 26.8 | Helped |
| use GT masks as the proposal gallery | PRM instance segmentation · PASCAL VOC 2012 val. | mAP^r_0.75: 9.0 -> 73.3 | Helped |
Closest papers by meaning
- Learning Instance Activation Maps for Weakly Supervised Instance Segmentation (CVPR 2019 · similarity 0.92)
- Weakly Supervised Learning of Instance Segmentation With Inter-Pixel Relations (CVPR 2019 · similarity 0.90)
- WISH: Weakly Supervised Instance Segmentation using Heterogeneous Labels (CVPR 2025 · similarity 0.89)
- Extreme Point Supervised Instance Segmentation (CVPR 2024 · similarity 0.89)
- P2Seg: Pointly-supervised Segmentation via Mutual Distillation (ICLR 2024 · similarity 0.88)
- Training-Free Dataset Pruning for Instance Segmentation (ICLR 2025 · 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.