Research map / Detection and segmentation / Weakly supervised detection and segmentation
Learning Saliency-Free Model with Generic Features for Weakly-Supervised Semantic Segmentation
Read the paperCheck 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 lower-layer features C42 instead of C43 for class patterns (ablation) | DCP proxy ground-truth · PASCAL VOC 2012 val | mIoU: 46.0 -> 52.4 | Helped |
| use lowest-layer features C41 instead of C42 (ablation) | DCP proxy ground-truth · PASCAL VOC 2012 val | mIoU: 52.4 -> 60.3 | Helped |
| retrain segmentation model on masks predicted by the first model (fills uncertain pixels) (ablation) | DCP with C41 · PASCAL VOC 2012 val | mIoU: 60.3 -> 64.5 | Helped |
| retrain a second time (ablation) | DCP with C41 · PASCAL VOC 2012 val | mIoU | No clear effect |
| replace CAM-based seeds plus saliency with saliency-free class patterns from C41 (ResNet101) | weakly-supervised segmentation · COCO val | mean IoU: 26.0 -> 29.9 | Helped |
| use ResNet101 instead of VGG16 backbone for segmentation network | Ours · PASCAL VOC 2012 val | mIoU: 61.2 -> 64.5 | Helped |
Closest papers by meaning
- Railroad Is Not a Train: Saliency As Pseudo-Pixel Supervision for Weakly Supervised Semantic Segmentation (CVPR 2021 · similarity 0.91)
- Combining Bottom-Up, Top-Down, and Smoothness Cues for Weakly Supervised Image Segmentation (CVPR 2017 · similarity 0.90)
- Learning Integral Objects With Intra-Class Discriminator for Weakly-Supervised Semantic Segmentation (CVPR 2020 · similarity 0.90)
- Joint Semantic Mining for Weakly Supervised RGB-D Salient Object Detection (NeurIPS 2021 · similarity 0.89)
- Exploiting Saliency for Object Segmentation From Image Level Labels (CVPR 2017 · similarity 0.89)
- 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.