From Pixel to Mask: A Survey of Out-of-Distribution Segmentation
Not on the map. Why?
Abstract (arXiv)
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.
Key facts. Cited by 1 paper in the map (data through 14 Oct 2026). Builds on 49 of its 66 references that are in the map. First posted on arXiv 14 Aug 2025.
Builds on
49 of 66 references.
References · 49
- 2021Learning Transferable Visual Models From Natural Language Supervisioncited by 33,744
- 2014ImageNet Large Scale Visual Recognition Challengecited by 8,498
- 2023Segment Anythingcited by 8,234
- 2023DINOv2: Learning Robust Visual Features without Supervisioncited by 7,974
- 2021Emerging Properties in Self-Supervised Vision Transformerscited by 6,968
- 2022Generative Adversarial Networkscited by 5,627
- 2016Image-to-Image Translation with Conditional Adversarial Networkscited by 4,527
- 2016The Cityscapes Dataset for Semantic Urban Scene Understandingcited by 4,444
- 2016Simple and Scalable Predictive Uncertainty Estimation using Deep Ensemblescited by 3,926
- 2023Grounding DINO: Marrying DINO with Grounded Pre-Training for Open-Set Object Detectioncited by 3,332
- 2016Perceptual Losses for Real-Time Style Transfer and Super-Resolutioncited by 2,694
- 2016A Baseline for Detecting Misclassified and Out-of-Distribution Examples in Neural Networkscited by 2,385
37 more
- 2021Masked-attention Mask Transformer for Universal Image Segmentationcited by 2,351
- 2018A Simple Unified Framework for Detecting Out-of-Distribution Samples and Adversarial Attackscited by 1,485
- 2017Enhancing The Reliability of Out-of-distribution Image Detection in Neural Networkscited by 1,328
- 2017High-Resolution Image Synthesis and Semantic Manipulation with Conditional GANscited by 1,179
- 2018Deep Anomaly Detection with Outlier Exposurecited by 942
- 2021ReAct: Out-of-distribution Detection With Rectified Activationscited by 486
- 2019Scaling Out-of-Distribution Detection for Real-World Settingscited by 465
- 2022ViM: Out-Of-Distribution with Virtual-logit Matchingcited by 405
- 2016Generating Images with Perceptual Similarity Metrics based on Deep Networkscited by 222
- 2023Robust Test-Time Adaptation in Dynamic Scenarioscited by 189
- 2021SegmentMeIfYouCan: A Benchmark for Anomaly Segmentationcited by 150
- 2019Detecting the Unexpected via Image Resynthesiscited by 137
- 2020Entropy Maximization and Meta Classification for Out-Of-Distribution Detection in Semantic Segmentationcited by 133
- 2019The Fishyscapes Benchmark: Measuring Blind Spots in Semantic Segmentationcited by 122
- 2016Lost and Found: Detecting Small Road Hazards for Self-Driving Vehiclescited by 117
- 2019Learning to Predict Layout-to-image Conditional Convolutions for Semantic Image Synthesiscited by 108
- 2020Synthesize then Compare: Detecting Failures and Anomalies for Semantic Segmentationcited by 96
- 2021Pixel-wise Energy-biased Abstention Learning for Anomaly Segmentation on Complex Urban Driving Scenescited by 83
- 2021Standardized Max Logits: A Simple yet Effective Approach for Identifying Unexpected Road Obstacles in Urban-Scene Segmentationcited by 77
- 2022RbA: Segmenting Unknown Regions Rejected by Allcited by 67
- 2022Residual Pattern Learning for Pixel-wise Out-of-Distribution Detection in Semantic Segmentationcited by 55
- 2018Prediction Error Meta Classification in Semantic Segmentation: Detection via Aggregated Dispersion Measures of Softmax Probabilitiescited by 54
- 2019Simultaneous Semantic Segmentation and Outlier Detection in Presence of Domain Shiftcited by 54
- 2018Discriminative out-of-distribution detection for semantic segmentationcited by 47
- 2023Unmasking Anomalies in Road-Scene Segmentationcited by 45
- 2023On Advantages of Mask-level Recognition for Outlier-aware Segmentationcited by 42
- 2023Towards Open-Set Test-Time Adaptation Utilizing the Wisdom of Crowds in Entropy Minimizationcited by 39
- 2023On the Robustness of Open-World Test-Time Training: Self-Training with Dynamic Prototype Expansioncited by 34
- 2019This is not what I imagined: Error Detection for Semantic Segmentation through Visual Dissimilaritycited by 16
- 2023UGainS: Uncertainty Guided Anomaly Instance Segmentationcited by 15
- 2023CSL: Class-Agnostic Structure-Constrained Learning for Segmentation Including the Unseencited by 15
- 2024STAMP: Outlier-Aware Test-Time Adaptation with Stable Memory Replaycited by 13
- 2023Have We Ever Encountered This Before? Retrieving Out-of-Distribution Road Obstacles from Driving Scenescited by 9
- 2023Calibrated Out-of-Distribution Detection with a Generic Representationcited by 8
- 2024Evaluation of Out-of-Distribution Detection Performance on Autonomous Driving Datasetscited by 5
- 2024Segment-Level Road Obstacle Detection Using Visual Foundation Model Priors and Likelihood Ratioscited by 3
- 2025Benchmarking Vision Foundation Models for Input Monitoring in Autonomous Drivingcited by 2
Built on by
Citing papers per year: 2026: 1.
1 paper.