Papers
Topics
Authors
Recent
Gemini 2.5 Flash
Gemini 2.5 Flash
169 tokens/sec
GPT-4o
7 tokens/sec
Gemini 2.5 Pro Pro
45 tokens/sec
o3 Pro
4 tokens/sec
GPT-4.1 Pro
38 tokens/sec
DeepSeek R1 via Azure Pro
28 tokens/sec
2000 character limit reached

Criteria for Uncertainty-based Corner Cases Detection in Instance Segmentation (2404.11266v1)

Published 17 Apr 2024 in cs.CV

Abstract: The operating environment of a highly automated vehicle is subject to change, e.g., weather, illumination, or the scenario containing different objects and other participants in which the highly automated vehicle has to navigate its passengers safely. These situations must be considered when developing and validating highly automated driving functions. This already poses a problem for training and evaluating deep learning models because without the costly labeling of thousands of recordings, not knowing whether the data contains relevant, interesting data for further model training, it is a guess under which conditions and situations the model performs poorly. For this purpose, we present corner case criteria based on the predictive uncertainty. With our corner case criteria, we are able to detect uncertainty-based corner cases of an object instance segmentation model without relying on ground truth (GT) data. We evaluated each corner case criterion using the COCO and the NuImages dataset to analyze the potential of our approach. We also provide a corner case decision function that allows us to distinguish each object into True Positive (TP), localization and/or classification corner case, or False Positive (FP). We also present our first results of an iterative training cycle that outperforms the baseline and where the data added to the training dataset is selected based on the corner case decision function.

Definition Search Book Streamline Icon: https://streamlinehq.com
References (43)
  1. Christopher M. Bishop. Pattern Recognition and Machine Learning (Information Science and Statistics). Springer-Verlag, New York, NY, USA, 2006.
  2. Variational Inference for Dirichlet Process Mixtures. Bayesian Analysis, 1(1):121–144, 2006.
  3. Description of Corner Cases in Automated Driving: Goals and Challenges. In Proc. of ICCV, ERCVAD Workshop, pages 1023–1028, Virtual, 2021.
  4. Towards Corner Case Detection for Autonomous Driving. In Proc. of IV, pages 438–445, Paris, France, 2019.
  5. TIDE: A General Toolbox for Identifying Object Detection Errors. In Proc. of ECCV, pages 558–573, Virtual, 2020.
  6. What Does Really Count? Estimating Relevance of Corner Cases for Semantic Segmentation in Automated Driving. In Proc. of ICCV, Workshops, pages 3991–4000, Paris, France, 2023.
  7. nuScenes: A Multimodal Dataset for Autonomous Driving. In Proc. of CVPR, pages 11618–11628, Seattle, WA, USA, 2020.
  8. Entropy Maximization and Meta Classification for Out-of-Distribution Detection in Semantic Segmentation. In Proc. of ICCV, pages 5108–5117, Montreal, QC, Canada, 2021.
  9. Anomaly Detection: A Survey. ACM Computing Surveys, 41(3):1–58, 2009.
  10. A Review and Comparative Study on Probabilistic Object Detection in Autonomous Driving. IEEE Trans. on ITS, pages 1–20, 2021.
  11. Dropout as a Bayesian Approximation: Representing Model Uncertainty in Deep Learning. In Proc. of ICML, pages 1050–1059, New York, NY, USA, 2016.
  12. A Survey of Uncertainty in Deep Neural Networks. arXiv preprint arXiv:2107.03342, 2022.
  13. Explaining and Harnessing Adversarial Examples. In Proc. of ICLR, pages 1–10, San Diego, CA, USA, 2015.
  14. Novelty Detection in Continuously Changing Environments. Future Generation Computer Systems, 114:138–154, 2021.
  15. Probabilistic Object Detection: Definition and Evaluation. In Proc. of WACV, pages 1031–1040, Snowmass Village, CO, USA, 2020.
  16. BayesOD: A Bayesian Approach for Uncertainty Estimation in Deep Object Detectors. arXiv preprint arXiv:1903.03838, 2019.
  17. D. M. Hawkins. Identification of Outliers. Springer, Dordrecht, 1980.
  18. Mask R-CNN. In Proc. of ICCV, pages 2980–2988, Venice, Italy, 2017.
  19. An Application-Driven Conceptualization of Corner Cases for Perception in Highly Automated Driving. In Proc. of the IV, pages 644–651, Nagoya, Japan, 2021a.
  20. Towards Corner Case Detection by Modeling the Uncertainty of Instance Segmentation Networks. In Proc. of ICPR, IADS Workshop, pages 361–374, Milan, Italy, 2021b.
  21. Sampling-based Uncertainty Estimation for an Instance Segmentation Network. arXiv preprint arXiv:2305.14977, 2023.
  22. Inspect, Understand, Overcome: A Survey of Practical Methods for AI Safety. arXiv preprint arXiv:2104.14235, 2021.
  23. Aleatoric and Epistemic Uncertainty in Machine Learning: An Introduction to Concepts and Methods. Machine Learning, 110(3):457–506, 2021.
  24. TIVE: A Toolbox for Identifying Video Instance Segmentation Errors. arXiv preprint arXiv:2210.08856, 2022.
  25. Toward Optimal Probabilistic Active Learning Using a Bayesian Approach. Machine Learning, 110(6):1199––1231, 2021.
  26. Simple and Scalable Predictive Uncertainty Estimation Using Deep Ensembles. In Proc. of NIPS, pages 6402–6413, Long Beach, CA, USA, 2017.
  27. Imbalanced-learn: A Python Toolbox to Tackle the Curse of Imbalanced Datasets in Machine Learning. Journal of Machine Learning Research, 18(17):1–5, 2017.
  28. The Earth Mover’s distance is the Mallows distance: some insights from statistics. In Proc. of ICCV, pages 251–256, Vancouver, BC, Canada, 2001.
  29. Microsoft COCO: Common Objects in Context. arXiv preprint arXiv:1405.0312, 2015.
  30. Detecting the Unexpected via Image Resynthesis. In Proc. of ICCV, pages 2152–2161, Seoul, Korea (South), 2019.
  31. Uncertainty-aware Instance Segmentation using Dropout Sampling, 2019. Accessed 15-Feb.-2019.
  32. Frank Nielsen. On a Generalization of the Jensen–Shannon Divergence and the Jensen–Shannon Centroid. Entropy, 22(2):1–24, 2020.
  33. Out-of-Distribution Detection for Automotive Perception. In Proc. of ITSC, pages 2938–2943, Indianapolis, IN, USA, 2021.
  34. Corner Case Data Description and Detection. In Proc. of ICSE Workshop, pages 19–26, Madrid, Spain, 2021.
  35. Scikit-learn: Machine Learning in Python. JMLR, 12(85):2825–2830, 2011.
  36. On Why the System Makes the Corner Case: AI-based HolisticAnomaly Detection for Autonomous Driving. In Proc. of IV, pages 337–344, Achen, Germany, 2022.
  37. A Review of Novelty Detection. Signal Processing, 99:215–249, 2014.
  38. Deep Neural Networks and Data for AutomatedDriving: Robustness, Uncertainty Quantification, and In-sights Towards Safety, chapter Uncertainty Quantification for ObjectDetection: Output- and Gradient-BasedApproaches, pages 251–275. Springer International Publishing, Cham, Switzerland, 2022.
  39. Space, Time, and Interaction: A Taxonomy of Corner Cases in Trajectory Datasets for Automated Driving. In Proc. of SSCI, IEEE CIVTS, pages 1–8, Singapore, 2022.
  40. Uncertainty-Based Out-of-Distribution Detection in Deep Reinforcement Learning. arXiv preprint arXiv:1901.02219, 2019.
  41. End-to-End People Detection in Crowded Scenes. In Proc. of CVPR, pages 2325–2333, Las Vegas, NV, USA, 2016.
  42. Misbehaviour Prediction for Autonomous Driving Systems. In Proc. of ICSE, pages 359–371, Seoul, Korea (South), 2020.
  43. Synthesize then Compare: Detecting Failures and Anomalies for Semantic Segmentation. arXiv preprint arXiv:2003.08440, 2020.

Summary

We haven't generated a summary for this paper yet.

Ai Sparkles Streamline Icon: https://streamlinehq.com Summarize Now
X Twitter Logo Streamline Icon: https://streamlinehq.com

Tweets