Research map / Label-efficient and robust learning / Domain adaptation and generalization
Continuous Pseudo-Label Rectified Domain Adaptive Semantic Segmentation With Implicit Neural Representations
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 learned continuous IR2F-RMM rectification | HRDA, MiT-B5 · GTA → Cityscapes | mIoU: 73.8 -> 74.4 | Helped |
| Add learned continuous IR2F-RMM rectification | HRDA, MiT-B5 · SYNTHIA → Cityscapes | mIoU: 65.8 -> 67.7 | Helped |
| Add learned continuous IR2F-RMM rectification | HRDA, MiT-B5 · Cityscapes → Dark Zurich | mIoU: 55.9 -> 58.9 | Helped |
| Add learned continuous IR2F-RMM rectification | HRDA, MiT-B5 · Cityscapes → ACDC-Night | mIoU: 53.9 -> 57.3 | Helped |
| Add IR2F-RMM to MRNet | MRNet · GTA → Cityscapes | mIoU: 50.3 -> 52.3 | Helped |
| Replace learned rectification with equal mixture weights (ablation) | HRDA · GTA → Cityscapes | mIoU: 74.4 -> 71.0 | Hurt |
| Replace learned rectification with equal mixture weights (ablation) | HRDA · SYNTHIA → Cityscapes | mIoU: 67.7 -> 61.9 | Hurt |
| Replace IR2F with convolutional blocks without coordinate input (ablation) | HRDA · GTA → Cityscapes | mIoU: 74.4 -> 72.9 | Hurt |
| Use last-layer outputs instead of member features as IR2F inputs (ablation) | HRDA · GTA → Cityscapes | mIoU: 74.4 -> 73.3 | Hurt |
| Use an INR-based segmentation decoder head instead of pseudo-label rectification (ablation) | HRDA · GTA → Cityscapes | mIoU: 74.4 -> 73.1 | Hurt |
| Use prediction variance for pseudo-label rectification (ablation) | HRDA · GTA → Cityscapes | mIoU: 74.4 -> 73.7 | Hurt |
| Use MC-Dropout predictions for ensemble rectification (ablation) | HRDA · GTA → Cityscapes | mIoU: 74.4 -> 71.8 | Hurt |
Closest papers by meaning
- MetaCorrection: Domain-Aware Meta Loss Correction for Unsupervised Domain Adaptation in Semantic Segmentation (CVPR 2021 · similarity 0.91)
- Unsupervised Model Adaptation for Continual Semantic Segmentation (AAAI 2021 · similarity 0.91)
- Prototypical Pseudo Label Denoising and Target Structure Learning for Domain Adaptive Semantic Segmentation (CVPR 2021 · similarity 0.91)
- Balanced Learning for Domain Adaptive Semantic Segmentation (ICML 2025 · similarity 0.90)
- Universal Domain Adaptation for Semantic Segmentation (CVPR 2025 · similarity 0.90)
- Towards Unsupervised Domain Bridging via Image Degradation in Semantic Segmentation (NeurIPS 2025 · similarity 0.87)
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.