mlsanityData to 7 Oct 2026

Label-Free Model Failure Detection for Lidar-based Point Cloud Segmentation

arXiv 2407.14306 ↗First posted 19 Jul 2024cs.RO cs.AIcited by 1

Not on the map. Why?

Abstract (arXiv)

Autonomous vehicles drive millions of miles on the road each year. Under such circumstances, deployed machine learning models are prone to failure both in seemingly normal situations and in the presence of outliers. However, in the training phase, they are only evaluated on small validation and test sets, which are unable to reveal model failures due to their limited scenario coverage. While it is difficult and expensive to acquire large and representative labeled datasets for evaluation, large-scale unlabeled datasets are typically available. In this work, we introduce label-free model failure detection for lidar-based point cloud segmentation, taking advantage of the abundance of unlabeled data available. We leverage different data characteristics by training a supervised and self-supervised stream for the same task to detect failure modes. We perform a large-scale qualitative analysis and present LidarCODA, the first publicly available dataset with labeled anomalies in real-world lidar data, for an extensive quantitative analysis.

Key facts. Cited by 1 paper in the map (data through 7 Oct 2026). Builds on 47 of its 78 references that are in the map. First posted on arXiv 19 Jul 2024.

Builds on

47 of 78 references.

References · 47

  1. 2019nuScenes: A multimodal dataset for autonomous drivingcited by 4,485
  2. 2016Simple and Scalable Predictive Uncertainty Estimation using Deep Ensemblescited by 3,888
  3. 2019Scalability in Perception for Autonomous Driving: Waymo Open Datasetcited by 2,436
  4. 2018BDD100K: A Diverse Driving Dataset for Heterogeneous Multitask Learningcited by 1,282
  5. 2021KITTI-360: A Novel Dataset and Benchmarks for Urban Scene Understanding in 2D and 3Dcited by 570
  6. 2019Confident Learning: Estimating Uncertainty in Dataset Labelscited by 411
  7. 2020SalsaNext: Fast, Uncertainty-aware Semantic Segmentation of LiDAR Point Clouds for Autonomous Drivingcited by 226
  8. 2021One Million Scenes for Autonomous Driving: ONCE Datasetcited by 218
  9. 2019Improving Adversarial Robustness via Promoting Ensemble Diversitycited by 213
  10. 2020Adversarial Robustness: From Self-Supervised Pre-Training to Fine-Tuningcited by 152
  11. 2020Are Labels Always Necessary for Classifier Accuracy Evaluation?cited by 104
  12. 2022CODA: A Real-World Road Corner Case Dataset for Object Detection in Autonomous Drivingcited by 100
35 more
  1. 2019Just Go with the Flow: Self-Supervised Scene Flow Estimationcited by 98
  2. 2022Agreement-on-the-Line: Predicting the Performance of Neural Networks under Distribution Shiftcited by 94
  3. 2022Anomaly Detection in Autonomous Driving: A Surveycited by 87
  4. 2019Identifying Unknown Instances for Autonomous Drivingcited by 73
  5. 2020FlowStep3D: Model Unrolling for Self-Supervised Scene Flow Estimationcited by 71
  6. 2022Agree to Disagree: Diversity through Disagreement for Better Transferabilitycited by 69
  7. 2021Moving Object Segmentation in 3D LiDAR Data: A Learning-based Approach Exploiting Sequential Datacited by 64
  8. 2022"Curse of rarity" for autonomous vehiclescited by 62
  9. 2021Revisiting Uncertainty-based Query Strategies for Active Learning with Transformerscited by 59
  10. 2021Detecting Errors and Estimating Accuracy on Unlabeled Data with Self-training Ensemblescited by 54
  11. 2022Predicting Out-of-Distribution Error with the Projection Normcited by 50
  12. 2021SODA10M: A Large-Scale 2D Self/Semi-Supervised Object Detection Dataset for Autonomous Drivingcited by 43
  13. 2020Self-Supervised Learning of Non-Rigid Residual Flow and Ego-Motioncited by 41
  14. 2021Description of Corner Cases in Automated Driving: Goals and Challengescited by 35
  15. 2023Unsupervised Universal Image Segmentationcited by 34
  16. 2024Outlier detection by ensembling uncertainty with negative objectnesscited by 34
  17. 2021An Application-Driven Conceptualization of Corner Cases for Perception in Highly Automated Drivingcited by 33
  18. 2021Open-set 3D Object Detectioncited by 29
  19. 2021Corner Cases for Visual Perception in Automated Driving: Some Guidance on Detection Approachescited by 27
  20. 2024LISO: Lidar-only Self-Supervised 3D Object Detectioncited by 22
  21. 2024UNION: Unsupervised 3D Object Detection using Object Appearance-based Pseudo-Classescited by 22
  22. 2024Multiagent Multitraversal Multimodal Self-Driving: Open MARS Datasetcited by 18
  23. 2023Perception Datasets for Anomaly Detection in Autonomous Driving: A Surveycited by 16
  24. 2023Identifying Label Errors in Object Detection Datasets by Loss Inspectioncited by 16
  25. 2024Energy-based Automated Model Evaluationcited by 16
  26. 2020Self-supervised Learning of LiDAR Odometry for Robotic Applicationscited by 15
  27. 2020YOdar: Uncertainty-based Sensor Fusion for Vehicle Detection with Camera and Radar Sensorscited by 13
  28. 2021Self-supervised and Supervised Joint Training for Resource-rich Machine Translationcited by 12
  29. 2018Informed Democracy: Voting-based Novelty Detection for Action Recognitioncited by 11
  30. 2022Learning-From-Disagreement: A Model Comparison and Visual Analytics Frameworkcited by 9
  31. 2022Active Learning by Query by Committee with Robust Divergencescited by 9
  32. 2022Improving robustness and calibration in ensembles with diversity regularizationcited by 8
  33. 2022Combating noisy labels in object detection datasetscited by 7
  34. 2024LaneCorrect: Self-supervised Lane Detectioncited by 5
  35. 2022Safe Perception -- A Hierarchical Monitor Approachcited by 1

Built on by

Cited by year (2026 to 7 Oct)

Citing papers per year: 2026: 1.

1 paper.

  1. 2026Online Monitoring Framework for Automotive Time Series Data using JEPA Embeddingscited by 0