From Pixel to Mask: A Survey of Out-of-Distribution Segmentation
Fuente:
arXiv
Salvato in:
| Autori principali: | , , |
|---|---|
| Natura: | Preprint |
| Pubblicazione: |
2025
|
| Soggetti: | |
| Accesso online: | |
| Tags: |
Aggiungi Tag
Nessun Tag, puoi essere il primo ad aggiungerne!!
|
| _version_ | 1866916898410594304 |
|---|---|
| author | Zhao, Wenjie Li, Jia Guo, Yunhui |
| author_facet | Zhao, Wenjie Li, Jia Guo, Yunhui |
| contents | Out-of-distribution (OoD) detection and segmentation have attracted growing attention as concerns about AI security rise. Conventional OoD detection methods identify the existence of OoD objects but lack spatial localization, limiting their usefulness in downstream tasks. OoD segmentation addresses this limitation by localizing anomalous objects at pixel-level granularity. This capability is crucial for safety-critical applications such as autonomous driving, where perception modules must not only detect but also precisely segment OoD objects, enabling targeted control actions and enhancing overall system robustness. In this survey, we group current OoD segmentation approaches into four categories: (i) test-time OoD segmentation, (ii) outlier exposure for supervised training, (iii) reconstruction-based methods, (iv) and approaches that leverage powerful models. We systematically review recent advances in OoD segmentation for autonomous-driving scenarios, identify emerging challenges, and discuss promising future research directions. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2508_10309 |
| institution | arXiv |
| publishDate | 2025 |
| record_format | arxiv |
| spellingShingle | From Pixel to Mask: A Survey of Out-of-Distribution Segmentation Zhao, Wenjie Li, Jia Guo, Yunhui Computer Vision and Pattern Recognition Out-of-distribution (OoD) detection and segmentation have attracted growing attention as concerns about AI security rise. Conventional OoD detection methods identify the existence of OoD objects but lack spatial localization, limiting their usefulness in downstream tasks. OoD segmentation addresses this limitation by localizing anomalous objects at pixel-level granularity. This capability is crucial for safety-critical applications such as autonomous driving, where perception modules must not only detect but also precisely segment OoD objects, enabling targeted control actions and enhancing overall system robustness. In this survey, we group current OoD segmentation approaches into four categories: (i) test-time OoD segmentation, (ii) outlier exposure for supervised training, (iii) reconstruction-based methods, (iv) and approaches that leverage powerful models. We systematically review recent advances in OoD segmentation for autonomous-driving scenarios, identify emerging challenges, and discuss promising future research directions. |
| title | From Pixel to Mask: A Survey of Out-of-Distribution Segmentation |
| topic | Computer Vision and Pattern Recognition |
| url | https://arxiv.org/abs/2508.10309 |