Research map / Detection and segmentation / Open-set and few-shot detection
PFENet++: Boosting Few-Shot Semantic Segmentation With the Noise-Filtered Context-Aware Prior Mask
Xiaoliu Luo, Zhuotao Tian, Taiping Zhang, Bei Yu, Yuan Yan Tang, Jiaya Jia
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 NSM to PFENet prior mask (ablation) | PFENet · PASCAL-5i (ResNet-50) | mIoU 1-shot/5-shot: 61.3 -> 63.2 (1-shot), 63.0 -> 66.8 (5-shot) | Improved |
| add CAPM {1,3,5} without NSM (ablation) | PFENet · PASCAL-5i (ResNet-50) | mIoU 1-shot/5-shot: 61.3 -> 58.6 (1-shot), 63.0 -> 59.7 (5-shot) | Hurt |
| add NSM on top of CAPM {1,3,5} (ablation) | CAPM · PASCAL-5i (ResNet-50) | mIoU 1-shot/5-shot: 58.6 -> 64.9 (1-shot), 59.7 -> 69.9 (5-shot) | Improved |
| multi-patch CAPM {1,3,5} instead of single patch {1} (ablation) | PFENet++ with NSM · PASCAL-5i (ResNet-50) | mIoU 1-shot/5-shot: 63.2 -> 64.9 (1-shot), 66.8 -> 69.9 (5-shot) | Improved |
| add patch size 7 to CAPM {1,3,5} (ablation) | PFENet++ · PASCAL-5i (ResNet-50) | mIoU 1-shot/5-shot: 64.9 -> 64.8 (1-shot), 69.9 -> 69.4 (5-shot) | No clear effect |
| replace CAPM with PPM after NSM (ablation) | PFENet++ · PASCAL-5i (ResNet-50) | mIoU 1-shot: 64.9 -> 51.3 (1-shot), 69.9 -> 53.2 (5-shot) | Hurt |
| replace CAPM with ASPP after NSM (ablation) | PFENet++ · PASCAL-5i (ResNet-50) | mIoU 1-shot: 64.9 -> 55.3 (1-shot), 69.9 -> 56.5 (5-shot) | Hurt |
| replace CAPM with GAU after NSM (ablation) | PFENet++ · PASCAL-5i (ResNet-50) | mIoU 1-shot: 64.9 -> 59.6 (1-shot), 69.9 -> 61.5 (5-shot) | Hurt |
| add middle-level prior masks with NSM (ablation) | PFENet++ · PASCAL-5i (ResNet-50) | mIoU 1-shot/5-shot: 64.1 -> 64.9 (1-shot), 68.6 -> 69.9 (5-shot) | Improved |
| add middle-level prior masks without NSM (ablation) | PFENet · PASCAL-5i (ResNet-50) | mIoU 1-shot/5-shot: 61.3 -> 61.0 (1-shot), 63.0 -> 63.1 (5-shot) | No clear effect |
| add 1x1 channel compression 2048 to 256 (ablation) | PFENet++ · PASCAL-5i (ResNet-50) | mIoU and FPS: 62.2 -> 63.1 (1-shot), 3.8 -> 9.1 FPS | Improved |
| add 2x2 average pooling on support features (ablation) | PFENet++ · PASCAL-5i (ResNet-50) | mIoU 1-shot/5-shot: 63.1 -> 64.9 (1-shot), 66.7 -> 69.9 (5-shot), 9.1 -> 12.7 FPS | Improved |
Later papers that compared with this method
1 later accepted paper reports PFENet as a baseline in their results tables; 1 of them report a better number on most of the results they share with it.
- Harmonizing Base and Novel Classes: A Class-Contrastive Approach for Generalized Few-Shot Segmentation (IJCV 2023 · better on 2 of 2 shared results)
Matched by method name in results tables read by AI. Authors' own tables tend to favour the newer method, and a comparison can be missed when a method is renamed.
Closest papers by meaning
- Relevant Intrinsic Feature Enhancement Network for Few-Shot Semantic Segmentation (AAAI 2024 · similarity 0.87)
- Rethinking Prior Information Generation with CLIP for Few-Shot Segmentation (CVPR 2024 · similarity 0.87)
- Bridge the Points: Graph-based Few-shot Segment Anything Semantically (NeurIPS 2024 · similarity 0.87)
- SANSA: Unleashing the Hidden Semantics in SAM2 for Few-Shot Segmentation (NeurIPS 2025 · similarity 0.87)
- Unlocking the Potential of Pre-trained Vision Transformers for Few-Shot Semantic Segmentation through Relationship Descriptors (CVPR 2024 · similarity 0.86)
- Unleashing the Potential of the Diffusion Model in Few-shot Semantic Segmentation (NeurIPS 2024 · similarity 0.86)
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.
