Research map / Detection and segmentation / Weakly supervised detection and segmentation
Causal Intervention for Weakly-Supervised Semantic Segmentation
Dong Zhang, Hanwang Zhang, Jinhui Tang, Xian-Sheng Hua, Qianru Sun
Read the paperCheck your idea against 137,217 papers
AI-extracted. The measured effects below were read by AI from the paper's full text (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 |
|---|---|---|---|
| Apply context adjustment to SEAM | SEAM, ResNet-38 · PASCAL VOC 2012 val | mIoU: 64.5 -> 66.1 | Improved |
| Apply context adjustment to SEAM | SEAM, ResNet-38 · PASCAL VOC 2012 test | mIoU: 65.7 -> 66.7 | Improved |
| Apply context adjustment to IRNet | IRNet, ResNet-50 · MS-COCO val | mIoU: 32.6 -> 33.4 | Improved |
| Apply context adjustment to SEC | SEC, VGG-16 · MS-COCO val | mIoU: 22.4 -> 23.7 | Improved |
| Replace M_t with the segmentation mask (ablation) | CONTA on SEAM · PASCAL VOC 2012 val | Seg. Mask mIoU: 64.3 -> 64.0 | Hurt |
| Use round 3 rather than round 1 (ablation) | CONTA on SEAM · PASCAL VOC 2012 val | Seg. Mask mIoU: 65.0 -> 66.1 | Improved |
| Continue from round 3 to round 4 (ablation) | CONTA on SEAM · PASCAL VOC 2012 val | Seg. Mask mIoU: 66.1 -> 65.5 | Hurt |
| Concatenate M_t at block-5 instead of block-2 (ablation) | CONTA on SEAM · PASCAL VOC 2012 val | Seg. Mask mIoU: 65.2 -> 66.1 | Improved |
| Construct the confounder set from segmentation masks instead of pseudo-masks (ablation) | CONTA on SEAM · PASCAL VOC 2012 val | Seg. Mask mIoU: 65.8 -> 66.1 | Improved |
| Apply CONTA across different WSSS baselines | SEC, SEAM, IRNet, and DSRG · PASCAL VOC 2012 | Segmentation-mask mIoU: +2.5; +1.8; +2.3; +1.4 | Improved |
Closest papers by meaning
- Progressive Feature Self-Reinforcement for Weakly Supervised Semantic Segmentation (AAAI 2024 · similarity 0.90)
- Threshold Matters in WSSS: Manipulating the Activation for the Robust and Accurate Segmentation Model Against Thresholds (CVPR 2022 · similarity 0.89)
- W2P: Switching from Weak Supervision to Partial Supervision for Semantic Segmentation (AAAI 2024 · similarity 0.89)
- Reliability Does Matter: An End-to-End Weakly Supervised Semantic Segmentation Approach (AAAI 2020 · similarity 0.89)
- Separate and Conquer: Decoupling Co-occurrence via Decomposition and Representation for Weakly Supervised Semantic Segmentation (CVPR 2024 · similarity 0.88)
- From SAM to CAMs: Exploring Segment Anything Model for Weakly Supervised Semantic Segmentation (CVPR 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.
