Research map / Detection and segmentation / Open-set and few-shot detection
OmniSAM: Omnidirectional Segment Anything Model for UDA in Panoramic Semantic Segmentation
Ding Zhong, Xu Zheng, Chenfei Liao, Yuanhuiyi Lyu, Jialei Chen, Shengyang Wu, Linfeng Zhang, Xuming Hu
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 self-supervised pseudo-label loss (ablation) | OmniSAM-B · CS13-to-DP13 | mIoU: 56.49 -> 57.06 | Improved |
| add FPA loss on top of pseudo-label loss (static labels) (ablation) | OmniSAM-B · CS13-to-DP13 | mIoU: 57.06 -> 58.29 | Improved |
| add dynamic pseudo-label update (no FPA) (ablation) | OmniSAM-B · CS13-to-DP13 | mIoU: 57.06 -> 60.60 | Improved |
| add FPA loss to full config with dynamic updates (ablation) | OmniSAM-B · CS13-to-DP13 | mIoU: 60.60 -> 62.46 | Improved |
| add self-supervised pseudo-label loss (ablation) | OmniSAM-L · SPin8-to-SPan8 | mIoU: 76.85 -> 77.36 | Improved |
| add FPA loss without pseudo-label update (ablation) | OmniSAM-L · SPin8-to-SPan8 | mIoU: 77.36 -> 77.10 | No clear effect |
| add dynamic pseudo-label update (no FPA) (ablation) | OmniSAM-L · SPin8-to-SPan8 | mIoU: 77.36 -> 79.06 | Improved |
| add FPA loss to dynamic-update config (ablation) | OmniSAM-L · SPin8-to-SPan8 | mIoU: 79.06 -> 77.79 | Hurt |
| add FPA loss weight lambda=0.1 (vs 0) (ablation) | OmniSAM-B · CS13-to-DP13 | mIoU: 56.49 -> 62.46 | Improved |
| raise lambda from 0.1 to 1.0 (ablation) | OmniSAM-B · CS13-to-DP13 | mIoU: 62.46 -> 60.90 | Hurt |
| add memory attention (ablation) | OmniSAM-L · SPin8-to-SPan8 | mIoU: 78.02 -> 79.06 | Improved |
| add memory attention (ablation) | OmniSAM-B · SPin8-to-SPan8 | mIoU: 73.09 -> 74.72 | Improved |
Closest papers by meaning
- GoodSAM: Bridging Domain and Capacity Gaps via Segment Anything Model for Distortion-aware Panoramic Semantic Segmentation (CVPR 2024 · similarity 0.89)
- Semantics Distortion and Style Matter: Towards Source-free UDA for Panoramic Segmentation (CVPR 2024 · similarity 0.88)
- OmniSegmentor: A Flexible Multi-Modal Learning Framework for Semantic Segmentation (NeurIPS 2025 · similarity 0.87)
- Geometric Exploitation for Indoor Panoramic Semantic Segmentation (NeurIPS 2024 · similarity 0.86)
- OmniSeg3D: Omniversal 3D Segmentation via Hierarchical Contrastive Learning (CVPR 2024 · similarity 0.85)
- More than the Sum: Panorama-Language Models for Adverse Omni-Scenes (CVPR 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.
