Research map / Label-efficient and robust learning / Domain adaptation and generalization
Bridging Granularity Gaps: Hierarchical Semantic Learning for Cross-domain Few-shot Segmentation
Sujun Sun, Haowen Gu, Cheng Xie, Yanxu Ren, Mingwu Ren, Haofeng Zhang
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 |
|---|---|---|---|
| add Dual Style Randomization (ablation) | HSL ResNet-50 baseline · CD-FSS avg 1-shot | mIoU: 57.82 -> 60.44 | Improved |
| add Dual Style Randomization (ablation) | HSL ViT-B/16 baseline · CD-FSS avg 1-shot | mIoU: 62.24 -> 64.55 | Improved |
| add Hierarchical Semantic Mining (ablation) | HSL ResNet-50 baseline · CD-FSS avg 1-shot | mIoU: 57.82 -> 60.92 | Improved |
| add Hierarchical Semantic Mining (ablation) | HSL ViT-B/16 baseline · CD-FSS avg 1-shot | mIoU: 62.24 -> 65.29 | Improved |
| add PCMT on top of DSR+HSM (ablation) | HSL ViT-B/16 · CD-FSS avg 1-shot | mIoU: 67.05 -> 68.24 | Improved |
| add PCMT on top of DSR+HSM (ablation) | HSL ResNet-50 · CD-FSS avg 1-shot | mIoU: 62.97 -> 64.23 | Improved |
| use all four superpixel scales instead of only the finest (ablation) | HSL ViT-B/16 · CD-FSS avg 1-shot | mIoU: 65.43 -> 67.05 | Improved |
| use PCMT instead of OTSU threshold (ablation) | HSL ViT-B/16 · CD-FSS avg 1-shot | mIoU: 67.95 -> 68.24 | Improved |
| use OTSU threshold instead of fixed 0 (ablation) | HSL ViT-B/16 · ISIC 1-shot | mIoU: 55.79 -> 59.98 | Improved |
| use OTSU threshold instead of fixed 0 (ablation) | HSL ViT-B/16 · FSS-1000 1-shot | mIoU: 81.93 -> 80.43 | Hurt |
| use HSL instead of DRA | ResNet-50 · CD-FSS avg 1-shot | mIoU: 60.86 -> 64.23 | Improved |
| use HSL instead of LoEC | ViT-B/16 · CD-FSS avg 5-shot | mIoU: 70.43 -> 72.32 | Improved |
Closest papers by meaning
- Adapter Naturally Serves as Decoupler for Cross-Domain Few-Shot Semantic Segmentation (ICML 2025 · similarity 0.90)
- Label-Efficient Few-Shot Semantic Segmentation with Unsupervised Meta-Training (AAAI 2024 · similarity 0.90)
- The Devil is in Low-Level Features for Cross-Domain Few-Shot Segmentation (CVPR 2025 · similarity 0.90)
- Cross-Domain Few-Shot Segmentation via Iterative Support-Query Correspondence Mining (CVPR 2024 · similarity 0.90)
- Adapt Before Comparison: A New Perspective on Cross-Domain Few-Shot Segmentation (CVPR 2024 · similarity 0.90)
- Lightweight Frequency Masker for Cross-Domain Few-Shot Semantic Segmentation (NeurIPS 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.
