Research map / Detection and segmentation / CNN features
Renovate Yourself: Calibrating Feature Representation of Misclassified Pixels for 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 |
|---|---|---|---|
| Add RCH | DeepLabV3 with ResNet101 · Cityscapes val | mIoU: 79.78 -> 80.72 | Helped |
| Add RCH | HRNet-W48 · Cityscapes val | mIoU: 80.69 -> 81.83 | Helped |
| Add RCH | OCRNet with HRNet-W48 · Cityscapes val | mIoU: 81.11 -> 82.24 | Helped |
| Add RCH | SegFormer with MiT-B0 · Cityscapes val | mIoU: 76.2 -> 76.8 | Helped |
| Add RCH | SegFormer with MiT-B1 · Cityscapes val | mIoU: 78.5 -> 79.0 | Helped |
| Add RCH | SegFormer with MiT-B2 · Cityscapes val | mIoU: 81.0 -> 81.6 | Helped |
| Use learned prototypes instead of prototypes calculated from TP pixels (ablation) | OCRNet · Cityscapes train/val | mIoU: 81.75 -> 81.34 | Hurt |
| Calculate prototypes from ground-truth pixels rather than TP pixels (ablation) | OCRNet · Cityscapes train/val | mIoU: 81.75 -> 81.55 | Hurt |
| Add TP-derived prototypes instead of using no prototype (ablation) | OCRNet · Cityscapes train/val | mIoU: 81.11 -> 81.75 | Helped |
| Add FN compensation (ablation) | OCRNet with TP prototypes · Cityscapes train/val | mIoU: 81.75 -> 81.93 | Helped |
| Add FP penalty (ablation) | OCRNet with TP prototypes · Cityscapes train/val | mIoU: 81.75 -> 81.88 | Helped |
| Add both FN compensation and FP penalty (ablation) | OCRNet with TP prototypes · Cityscapes train/val | mIoU: 81.75 -> 82.24 | Helped |
Closest papers by meaning
- SFC: Shared Feature Calibration in Weakly Supervised Semantic Segmentation (AAAI 2024 · similarity 0.88)
- On Calibrating Semantic Segmentation Models: Analyses and an Algorithm (CVPR 2023 · similarity 0.88)
- Consistent-Separable Feature Representation for Semantic Segmentation (AAAI 2021 · similarity 0.88)
- MaSS13K: A Matting-level Semantic Segmentation Benchmark (CVPR 2025 · similarity 0.85)
- Improving Semi-Supervised Semantic Segmentation with Sliced-Wasserstein Feature Alignment and Uniformity (CVPR 2025 · similarity 0.85)
- Pixels to Tokens: Token Space Efficient Active Learning for Low-Budget Semantic Segmentation (NeurIPS 2026 · similarity 0.85)
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.