Research map / Detection and segmentation / Weakly supervised detection and segmentation
Weakly Supervised Semantic Segmentation Using Out-of-Distribution Data
Check 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 hard OoD training and metric learning to AdvCAM | AdvCAM with ResNet-101 · Pascal VOC 2012 val | mIoU: 67.5 -> 69.8 | Helped |
| Add hard OoD training and metric learning to AdvCAM | AdvCAM with ResNet-101 · Pascal VOC 2012 test | mIoU: 67.1 -> 69.9 | Helped |
| Add W-OoD to IRN | IRN · Pascal VOC 2012 train | mIoU: 49.5 -> 53.3 | Helped |
| Add W-OoD to SEAM | SEAM · Pascal VOC 2012 train | mIoU: 54.8 -> 55.9 | Helped |
| Add W-OoD to AdvCAM | AdvCAM · Pascal VOC 2012 train | precision: 66.8 -> 71.5 | Helped |
| Use classification loss to label hard OoD images as background (ablation) | W-OoD classifier · Pascal VOC 2012 train | mIoU: 49.5 -> 50.0 | Helped |
| Add metric loss on hard OoD data (ablation) | W-OoD classifier · Pascal VOC 2012 train | mIoU: 50.0 -> 50.2 -> 53.3 | Helped |
| Add metric loss on in-distribution data (ablation) | W-OoD classifier · Pascal VOC 2012 train | mIoU: 52.3 -> 53.3 | Helped |
| Use predicted-class clustering instead of K-means for OoD samples (ablation) | W-OoD · Pascal VOC 2012 train | mIoU: 52.1 vs 52.4 | No clear effect |
| Increase hard OoD image count from 0 to 20 (ablation) | W-OoD · Pascal VOC 2012 val | mIoU: 67.5 -> 68.1 | Helped |
| Increase hard OoD image count from 20 to 5190 (ablation) | W-OoD · Pascal VOC 2012 val | mIoU: 68.1 -> 69.8 | Helped |
| Use W-OoD rather than the baseline on dining table localization | W-OoD · Pascal VOC 2012 class-wise mIoU | mIoU | Hurt |
Closest papers by meaning
- Separate and Conquer: Decoupling Co-occurrence via Decomposition and Representation for Weakly Supervised Semantic Segmentation (CVPR 2024 · similarity 0.88)
- Progressive Feature Self-Reinforcement for Weakly Supervised Semantic Segmentation (AAAI 2024 · similarity 0.88)
- Weakly Supervised Semantic Segmentation for Driving Scenes (AAAI 2024 · similarity 0.87)
- Weakly Supervised Open-Vocabulary Object Detection (AAAI 2024 · similarity 0.87)
- Class Tokens Infusion for Weakly Supervised Semantic Segmentation (CVPR 2024 · similarity 0.87)
- Weakly Supervised Semantic Segmentation via Progressive Confidence Region Expansion (CVPR 2025 · 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.