mlsanityData to 7 Oct 2026

Dense Out-of-Distribution Detection by Robust Learning on Synthetic Negative Data

arXiv 2112.12833 ↗First posted 23 Dec 2021cs.CVcited by 10

Not on the map. Why?

Abstract (arXiv)

Standard machine learning is unable to accommodate inputs which do not belong to the training distribution. The resulting models often give rise to confident incorrect predictions which may lead to devastating consequences. This problem is especially demanding in the context of dense prediction since input images may be only partially anomalous. Previous work has addressed dense out-of-distribution detection by discriminative training with respect to off-the-shelf negative datasets. However, real negative data are unlikely to cover all modes of the entire visual world. To this end, we extend this approach by generating synthetic negative patches along the border of the inlier manifold. We leverage a jointly trained normalizing flow due to coverage-oriented learning objective and the capability to generate samples at different resolutions. We detect anomalies according to a principled information-theoretic criterion which can be consistently applied through training and inference. The resulting models set the new state of the art on benchmarks for out-of-distribution detection in road-driving scenes and remote sensing imagery, in spite of minimal computational overhead.

Key facts. Cited by 10 papers in the map (data through 7 Oct 2026). Builds on 63 of its 102 references that are in the map. First posted on arXiv 23 Dec 2021.

Builds on

63 of 102 references.

References · 63

  1. 2017On Calibration of Modern Neural Networkscited by 4,860
  2. 2016The Cityscapes Dataset for Semantic Urban Scene Understandingcited by 4,427
  3. 2016A Baseline for Detecting Misclassified and Out-of-Distribution Examples in Neural Networkscited by 2,362
  4. 2017What Uncertainties Do We Need in Bayesian Deep Learning for Computer Vision?cited by 2,335
  5. 2016Density estimation using Real NVPcited by 2,029
  6. 2018Glow: Generative Flow with Invertible 1x1 Convolutionscited by 1,539
  7. 2017Enhancing The Reliability of Out-of-distribution Image Detection in Neural Networkscited by 1,325
  8. 2020Energy-based Out-of-distribution Detectioncited by 1,296
  9. 2018Deep Anomaly Detection with Outlier Exposurecited by 939
  10. 2016Pixel Recurrent Neural Networkscited by 814
  11. 2020NVAE: A Deep Hierarchical Variational Autoencodercited by 635
  12. 2018Predictive Uncertainty Estimation via Prior Networkscited by 555
38 more
  1. 2015Towards Open Set Deep Networkscited by 533
  2. 2021ReAct: Out-of-distribution Detection With Rectified Activationscited by 486
  3. 2019Scaling Out-of-Distribution Detection for Real-World Settingscited by 465
  4. 2020A Unifying Review of Deep and Shallow Anomaly Detectioncited by 423
  5. 2017Training Confidence-calibrated Classifiers for Detecting Out-of-Distribution Samplescited by 389
  6. 2019Panoptic-DeepLabcited by 362
  7. 2021Open-Set Recognition: a Good Closed-Set Classifier is All You Need?cited by 351
  8. 2017One-Shot Learning for Semantic Segmentationcited by 337
  9. 2018Recent Advances in Open Set Recognition: A Surveycited by 309
  10. 2017A General and Adaptive Robust Loss Functioncited by 236
  11. 2016Unrolled Generative Adversarial Networkscited by 208
  12. 2018Reducing Network Agnostophobiacited by 165
  13. 2021Adversarial Reciprocal Points Learning for Open Set Recognitioncited by 153
  14. 2021SegmentMeIfYouCan: A Benchmark for Anomaly Segmentationcited by 150
  15. 2019Detecting the Unexpected via Image Resynthesiscited by 137
  16. 2021OpenGAN: Open-Set Recognition via Open Data Generationcited by 135
  17. 2018Improving Semantic Segmentation via Video Propagation and Label Relaxationcited by 134
  18. 2020Entropy Maximization and Meta Classification for Out-Of-Distribution Detection in Semantic Segmentationcited by 133
  19. 2019The Fishyscapes Benchmark: Measuring Blind Spots in Semantic Segmentationcited by 122
  20. 2016Lost and Found: Detecting Small Road Hazards for Self-Driving Vehiclescited by 117
  21. 2020Hybrid Models for Open Set Recognitioncited by 109
  22. 2021Pixel-wise Anomaly Detection in Complex Driving Scenescited by 109
  23. 2020Synthesize then Compare: Detecting Failures and Anomalies for Semantic Segmentationcited by 96
  24. 2021Understanding Failures in Out-of-Distribution Detection with Deep Generative Modelscited by 89
  25. 2021On Feature Collapse and Deep Kernel Learning for Single Forward Pass Uncertaintycited by 87
  26. 2017In-Place Activated BatchNorm for Memory-Optimized Training of DNNscited by 81
  27. 2021Standardized Max Logits: A Simple yet Effective Approach for Identifying Unexpected Road Obstacles in Urban-Scene Segmentationcited by 77
  28. 2019Simultaneous Semantic Segmentation and Outlier Detection in Presence of Domain Shiftcited by 54
  29. 2021Deep Metric Learning for Open World Semantic Segmentationcited by 51
  30. 2021Densely connected normalizing flowscited by 49
  31. 2018Discriminative out-of-distribution detection for semantic segmentationcited by 47
  32. 2020Robust Unsupervised Video Anomaly Detection by Multi-Path Frame Predictioncited by 42
  33. 2019Knowledge Distillation for Incremental Learning in Semantic Segmentationcited by 40
  34. 2019TRADI: Tracking deep neural network weight distributions for uncertainty estimationcited by 39
  35. 2018Predicting Future Instance Segmentation by Forecasting Convolutional Featurescited by 37
  36. 2020MOCCA: Multi-Layer One-Class ClassificAtion for Anomaly Detectioncited by 31
  37. 2020Dense open-set recognition with synthetic outliers generated by Real NVPcited by 29
  38. 2020Detecting Road Obstacles by Erasing Themcited by 29

50 most cited of 63 All 63 (JSON)

Built on by

Cited by year (2026 to 7 Oct)

Citing papers per year: 2022: 1; 2023: 1; 2024: 5; 2025: 2; 2026: 1.

10 papers.

  1. 2024Outlier detection by ensembling uncertainty with negative objectnesscited by 34
  2. 2024PixOOD: Pixel-Level Out-of-Distribution Detectioncited by 20
  3. 2024AnoVox: A Benchmark for Multimodal Anomaly Detection in Autonomous Drivingcited by 14
  4. 2024A Likelihood Ratio-Based Approach to Segmenting Unknown Objectscited by 12
  5. 2023Hybrid Open-set Segmentation with Synthetic Negative Datacited by 8
  6. 2022AttEntropy: On the Generalization Ability of Supervised Semantic Segmentation Transformers to New Objects in New Domainscited by 4
  7. 2024Segment-Level Road Obstacle Detection Using Visual Foundation Model Priors and Likelihood Ratioscited by 3
  8. 2025Uncertainty-Aware Likelihood Ratio Estimation for Pixel-Wise Out-of-Distribution Detectioncited by 3
  9. 2025Anomalies by Synthesis: Anomaly Detection using Generative Diffusion Models for Off-Road Navigationcited by 1
  10. 2026Real-World On-Vehicle Evaluation of Embedding-Based Anomaly Detectioncited by 0