Research map / Detection and segmentation / Weakly supervised detection and segmentation
Learning Integral Objects With Intra-Class Discriminator for Weakly-Supervised Semantic Segmentation
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 |
|---|---|---|---|
| Use intra-class foreground/background discrimination with ICD instead of CAM-only seeds | Image-level WSSS on VOC 2012 · VOC 2012 val | mIoU: 58.4 -> 61.2 | Helped |
| Add external saliency estimates for background seeds | ICD with ResNet101 · VOC 2012 test | mIoU: 64.3 -> 68.0 | Helped |
| Add superpixel refinement (ablation) | Bottom-up ICD seed estimation · VOC 2012 training seeds | mIoU: 49.9 -> 54.0 | Helped |
| Add top-down adaptation (ablation) | Bottom-up ICD with superpixel refinement · VOC 2012 training seeds | mIoU: 54.0 -> 59.9 | Helped |
| Add CRF post-processing (ablation) | ICD seeds with top-down adaptation · VOC 2012 training seeds | mIoU: 59.9 -> 62.2 | Helped |
| Replace Lbu with Labs (ablation) | ICD estimation · VOC 2012 training seeds | mIoU: 62.2 -> 58.3 | Hurt |
| Replace Lbu with Lsig (ablation) | ICD estimation · VOC 2012 training seeds | mIoU: 62.2 -> 52.9 | Hurt |
| Use Conv4 rather than Conv5 features for bottom-up estimation (ablation) | ICD · VOC 2012 training seeds | mIoU: 49.9 -> 43.8 | Hurt |
| Use Fc6 rather than Conv5 features for bottom-up estimation (ablation) | ICD · VOC 2012 training seeds | mIoU: 49.9 -> 40.0 | Hurt |
| Use concatenated features rather than Conv5 features for bottom-up estimation (ablation) | ICD · VOC 2012 training seeds | mIoU: 49.9 -> 41.3 | Hurt |
| Use top-down adaptation directly on Conv5 features (ablation) | ICD · VOC 2012 training seeds | mIoU: 59.9 -> 50.4 | Hurt |
| Share top-down adaptation parameters with the CAM branch (ablation) | ICD · VOC 2012 training seeds | mIoU: 59.9 -> 59.2 | Hurt |
Closest papers by meaning
- Embedded Discriminative Attention Mechanism for Weakly Supervised Semantic Segmentation (CVPR 2021 · similarity 0.91)
- Class Tokens Infusion for Weakly Supervised Semantic Segmentation (CVPR 2024 · similarity 0.91)
- Progressive Feature Self-Reinforcement for Weakly Supervised Semantic Segmentation (AAAI 2024 · similarity 0.90)
- From SAM to CAMs: Exploring Segment Anything Model for Weakly Supervised Semantic Segmentation (CVPR 2024 · similarity 0.89)
- W2P: Switching from Weak Supervision to Partial Supervision for Semantic Segmentation (AAAI 2024 · similarity 0.89)
- Separate and Conquer: Decoupling Co-occurrence via Decomposition and Representation for Weakly Supervised Semantic Segmentation (CVPR 2024 · similarity 0.89)
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.