Research map / Detection and segmentation / Weakly supervised detection and segmentation
Single-Stage Semantic Segmentation From Image Labels
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 CRF post-processing | Single-stage method · Pascal VOC 2012 val | mean IoU (%): 59.7 -> 62.7 | Helped |
| Use focal mask penalty with p=5, λ=0.1 instead of p=0, λ=0.1 (ablation) | Single-stage method · Pascal VOC 2012 val | IoU (%): 58.8 -> 60.2 | Helped |
| Remove the focal mask penalty (ablation) | Single-stage method · Pascal VOC 2012 val | IoU (%): 56.6 IoU | Hurt |
| Add Pixel-Adaptive Mask Refinement (ablation) | Single-stage method · Pascal VOC 2012 val | IoU (%): 31.8 -> 59.4 | Helped |
| Increase PAMR dilation range to 1-2-4-8-12-24 (ablation) | Single-stage method · Pascal VOC 2012 val | IoU (%): 50.6 -> 59.4 | Helped |
| Add Stochastic Gate (ablation) | Single-stage method · Pascal VOC 2012 val | IoU (%): 55.6 -> 59.8 | Helped |
| Use stochastic rather than deterministic feature mixing (ablation) | Single-stage method · Pascal VOC 2012 val | IoU (%): 57.5 -> 59.8 | Helped |
| Add Global Cue Injection to the gate (ablation) | Single-stage method · Pascal VOC 2012 val | IoU (%): 59.4 vs 59.8; 62.2 vs 60.9 | Mixed |
| Use 10 rather than 5 PAMR iterations (ablation) | Single-stage method · Pascal VOC 2012 val | IoU (%): 59.1 -> 59.4 | Helped |
| Increase PAMR iterations from 10 to 15 (ablation) | Single-stage method · Pascal VOC 2012 val | IoU (%): 59.4 -> 59.0 | Hurt |
Closest papers by meaning
- Reliability Does Matter: An End-to-End Weakly Supervised Semantic Segmentation Approach (AAAI 2020 · similarity 0.93)
- Incremental Learning in Semantic Segmentation From Image Labels (CVPR 2022 · similarity 0.91)
- Semi-Supervised Semantic Image Segmentation With Self-Correcting Networks (CVPR 2020 · similarity 0.91)
- Progressive Feature Self-Reinforcement for Weakly Supervised Semantic Segmentation (AAAI 2024 · similarity 0.91)
- DuPL: Dual Student with Trustworthy Progressive Learning for Robust Weakly Supervised Semantic Segmentation (CVPR 2024 · similarity 0.90)
- From SAM to CAMs: Exploring Segment Anything Model for Weakly Supervised Semantic Segmentation (CVPR 2024 · similarity 0.90)
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.