Research map / Detection and segmentation / Weakly supervised detection and segmentation
Embedded Discriminative Attention Mechanism 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 |
|---|---|---|---|
| Replace CAM-based responses with class-specific masks from the DA layer (ablation) | EDAM pseudo-label generation · PASCAL VOC 2012 pseudo-labels | mIoU: 47.55% -> 51.58% | Helped |
| Add collaborative multi-attention to the DA layer (ablation) | EDAM pseudo-label generation · PASCAL VOC 2012 pseudo-labels | mIoU: 51.58% -> 52.83% | Helped |
| Add dCRF refinement to DA + CMA pseudo-labels (ablation) | EDAM pseudo-label generation · PASCAL VOC 2012 pseudo-labels | mIoU: 52.83% -> 58.18% | Helped |
| Use multiple images rather than a single image for CMA (ablation) | EDAM CMA module · PASCAL VOC 2012 pseudo-labels | mIoU: 52.14% -> 52.83% | Helped |
| Use two random images instead of two similar-label images in CMA (ablation) | EDAM CMA module · PASCAL VOC 2012 pseudo-labels | mIoU: 52.83% -> 52.77% | No clear effect |
| Increase CMA input from two to three similar-label images (ablation) | EDAM CMA module · PASCAL VOC 2012 pseudo-labels | mIoU: 52.83% -> 52.77% | No clear effect |
| Increase CMA input from two to four similar-label images (ablation) | EDAM CMA module · PASCAL VOC 2012 pseudo-labels | mIoU: 52.83% -> 52.52% | Hurt |
| Add saliency-map fusion using threshold θ (ablation) | EDAM pseudo-label post-processing · PASCAL VOC 2012 pseudo-labels | mIoU: 58.18% -> 66.10% | Helped |
| Add suppression of low-confidence salient regions using α (ablation) | EDAM pseudo-label post-processing · PASCAL VOC 2012 pseudo-labels | mIoU: 66.10% -> 66.98% | Helped |
| Add pop-up of high-confidence regions using β (ablation) | EDAM pseudo-label post-processing · PASCAL VOC 2012 pseudo-labels | mIoU: 66.10% -> 67.22% | Helped |
| Combine thresholding, suppression using α, and pop-up using β (ablation) | EDAM pseudo-label post-processing · PASCAL VOC 2012 pseudo-labels | mIoU: 66.10% -> 68.11% | Helped |
| Train EDAM with extra Caltech-256 single-label images | EDAM · PASCAL VOC 2012 val and test | mIoU: 67.1 -> 72.0 (val); 67.2 -> 71.4 (test) | Helped |
Closest papers by meaning
- Threshold Matters in WSSS: Manipulating the Activation for the Robust and Accurate Segmentation Model Against Thresholds (CVPR 2022 · similarity 0.91)
- 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)
- MoRe: Class Patch Attention Needs Regularization for Weakly Supervised Semantic Segmentation (AAAI 2025 · similarity 0.88)
- From SAM to CAMs: Exploring Segment Anything Model for Weakly Supervised Semantic Segmentation (CVPR 2024 · similarity 0.88)
- SFC: Shared Feature Calibration in Weakly Supervised Semantic Segmentation (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.