Research map / Label-efficient and robust learning / Domain adaptation and generalization
Denoise and Align: Towards Source-Free UDA for Robust Panoramic Semantic Segmentation
Yaowen Chang, Zhen Cao, Xu Zheng, Xiaoxin Mi, Zhen Dong
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 |
|---|---|---|---|
| DAPASS vs previous source-free SOTA 360SFUDA++ | DAPASS B2 · Cityscapes-to-DensePASS | mIoU: 52.99 -> 55.04 | Improved |
| DAPASS vs 360SFUDA++ | DAPASS B2 · SPin-to-SPan | mIoU: 68.84 -> 70.38 | Improved |
| add Path A bi-level neighbor denoising (ablation) | PCGD · Cityscapes-to-DensePASS (B1) | mIoU: 47.82 -> 50.23 | Improved |
| add Path B class-balancing copy-paste (ablation) | PCGD · Cityscapes-to-DensePASS (B1) | mIoU: 48.35 -> 50.23 | Improved |
| add Path B copy-paste (ablation) | PCGD · SPin-to-SPan (B1) | mIoU: 63.12 -> 65.87 | Improved |
| add Path A (ablation) | PCGD · SPin-to-SPan (B2) | mIoU: 63.55 -> 67.32 | Improved |
| add CRAM to PCGD (ablation) | DAPASS · Cityscapes-to-DensePASS (B1) | mIoU: 50.23 -> 53.16 | Improved |
| add CRAM to PCGD (ablation) | DAPASS · SPin-to-SPan (B1) | mIoU: 65.87 -> 69.56 | Improved |
| vary top-P% from 1 to 20 (ablation) | DAPASS B2 · Cityscapes-to-DensePASS | mIoU: 54.52 to 55.04 | No clear effect |
| vary tau from 2,000 to 10,000 (ablation) | DAPASS B2 · SPin-to-SPan | mIoU: 69.35 to 70.38 | No clear effect |
| use unweighted pseudo-labels instead of source-only (ablation) | self-training · Cityscapes-to-DensePASS (B1) | mIoU: 36.43 -> 40.71 | Improved |
Closest papers by meaning
- 360SFUDA++: Towards Source-Free UDA for Panoramic Segmentation by Learning Reliable Category Prototypes (TPAMI 2024 · similarity 0.91)
- Bending Reality: Distortion-Aware Transformers for Adapting to Panoramic Semantic Segmentation (CVPR 2022 · similarity 0.91)
- Semantics Distortion and Style Matter: Towards Source-free UDA for Panoramic Segmentation (CVPR 2024 · similarity 0.91)
- GoodSAM: Bridging Domain and Capacity Gaps via Segment Anything Model for Distortion-aware Panoramic Semantic Segmentation (CVPR 2024 · similarity 0.90)
- Geometric Exploitation for Indoor Panoramic Semantic Segmentation (NeurIPS 2024 · similarity 0.89)
- AdapToPASS: Ambiguity-aware Adaptive Spherical Transformer for Panoramic Semantic Segmentation (NeurIPS 2026 · 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.
