Research map / Detection and segmentation / Open-set and few-shot detection
Mining Latent Classes for Few-Shot Segmentation
Lihe Yang, Wei Zhuo, Lei Qi, Yinghuan Shi, Yang Gao
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 foreground prototype rectification (ablation) | baseline · PASCAL-5i 1-shot (ResNet-50) | mIoU: 57.8 -> 58.9 | Improved |
| add background prototype rectification (ablation) | baseline · PASCAL-5i 1-shot (ResNet-50) | mIoU: 57.8 -> 59.3 | Improved |
| add latent class mining (ablation) | baseline · PASCAL-5i 1-shot (ResNet-50) | mIoU: 57.8 -> 59.8 | Improved |
| add latent class mining (ablation) | FG + BG · PASCAL-5i 1-shot (ResNet-50) | mIoU: 60.4 -> 62.1 | Improved |
| mine latent classes from the training set as unlabeled data (ablation) | global background rectification only · PASCAL-5i 1-shot (ResNet-50) | mIoU: 59.3 -> 62.1 | Improved |
| add remaining raw training images as unlabeled data (ablation) | mining from trainset · PASCAL-5i 1-shot (ResNet-50) | mIoU: 62.1 -> 63.6 | Improved |
| add ImageNet subset to unlabeled data (ablation) | mining from trainset · PASCAL-5i 1-shot (ResNet-50) | mIoU: 62.1 -> 62.8 | Improved |
| use latent class mining instead of SimCLR on training set (ablation) | few-shot segmentation baseline · PASCAL-5i 1-shot (ResNet-50) | mIoU: 56.9 -> 59.8 | Improved |
| add EMA of model parameters (ablation) | full method · PASCAL-5i 1-shot (ResNet-50) | mIoU: 61.4 -> 62.1 | Improved |
| add EMA of model parameters (ablation) | baseline · PASCAL-5i 1-shot (ResNet-50) | mIoU: 57.8 -> 57.9 | No clear effect |
| set K-Means clusters to 5 instead of 1 (ablation) | full method · PASCAL-5i 1-shot (ResNet-50) | mIoU: 60.4 -> 62.1 | Improved |
| increase K-Means clusters from 5 to 9 (ablation) | full method · PASCAL-5i 1-shot (ResNet-50) | mIoU: 62.1 -> 60.6 | Hurt |
Closest papers by meaning
- Adaptive FSS: A Novel Few-Shot Segmentation Framework via Prototype Enhancement (AAAI 2024 · similarity 0.92)
- Learning What Not To Segment: A New Perspective on Few-Shot Segmentation (CVPR 2022 · similarity 0.92)
- Cross-Domain Few-Shot Segmentation via Iterative Support-Query Correspondence Mining (CVPR 2024 · similarity 0.89)
- Label-Efficient Few-Shot Semantic Segmentation with Unsupervised Meta-Training (AAAI 2024 · similarity 0.89)
- Adapt Before Comparison: A New Perspective on Cross-Domain Few-Shot Segmentation (CVPR 2024 · similarity 0.88)
- Unlocking the Power of SAM 2 for Few-Shot Segmentation (ICML 2025 · 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.
