mlsanityData to 14 Oct 2026

ProOOD: Prototype-Guided Out-of-Distribution 3D Occupancy Prediction

arXiv 2604.01081 ↗First posted 1 Apr 2026cs.CV cs.LG cs.ROcited by 0

Not on the map. Why?

Abstract (arXiv)

3D semantic occupancy prediction is central to autonomous driving, yet current methods are vulnerable to long-tailed class bias and out-of-distribution (OOD) inputs, often overconfidently assigning anomalies to rare classes. We present ProOOD, a lightweight, plug-and-play method that couples prototype-guided refinement with training-free OOD scoring. ProOOD comprises (i) prototype-guided semantic imputation that fills occluded regions with class-consistent features, (ii) prototype-guided tail mining that strengthens rare-class representations to curb OOD absorption, and (iii) EchoOOD, which fuses local logit coherence with local and global prototype matching to produce reliable voxel-level OOD scores. Extensive experiments on five datasets demonstrate that ProOOD achieves state-of-the-art performance on both in-distribution 3D occupancy prediction and OOD detection. On SemanticKITTI, it surpasses baselines by +3.57% mIoU overall and +24.80% tail-class mIoU; on VAA-KITTI, it improves AuPRCr by +19.34 points, with consistent gains across benchmarks. These improvements yield more calibrated occupancy estimates and more reliable OOD detection in safety-critical urban driving. The source code is publicly available at https://github.com/7uHeng/ProOOD.

Key facts. No citing papers in the map yet (data through 14 Oct 2026). Builds on 89 of its 110 references that are in the map. First posted on arXiv 1 Apr 2026.

Builds on

89 of 110 references.

References · 89

  1. 2015Deep Residual Learning for Image Recognitioncited by 52,096
  2. 2019EfficientNet: Rethinking Model Scaling for Convolutional Neural Networkscited by 6,442
  3. 2022BEVFormer: Learning Bird's-Eye-View Representation from Multi-Camera Images via Spatiotemporal Transformerscited by 1,218
  4. 2022Planning-oriented Autonomous Drivingcited by 1,087
  5. 2020Lift, Splat, Shoot: Encoding Images From Arbitrary Camera Rigs by Implicitly Unprojecting to 3Dcited by 1,044
  6. 2018Generalized Cross Entropy Loss for Training Deep Neural Networks with Noisy Labelscited by 997
  7. 2019Scaling Out-of-Distribution Detection for Real-World Settingscited by 465
  8. 2023Tri-Perspective View for Vision-Based 3D Semantic Occupancy Predictioncited by 409
  9. 2021MonoScene: Monocular 3D Semantic Scene Completioncited by 363
  10. 2023Occ3D: A Large-Scale 3D Occupancy Prediction Benchmark for Autonomous Drivingcited by 332
  11. 2023SurroundOcc: Multi-Camera 3D Occupancy Prediction for Autonomous Drivingcited by 316
  12. 2023VoxFormer: Sparse Voxel Transformer for Camera-based 3D Semantic Scene Completioncited by 313
38 more
  1. 2023OccFormer: Dual-path Transformer for Vision-based 3D Semantic Occupancy Predictioncited by 293
  2. 2019Addressing Failure Prediction by Learning Model Confidencecited by 186
  3. 2024GaussianFormer: Scene as Gaussians for Vision-Based 3D Semantic Occupancy Predictioncited by 157
  4. 2018Evaluating Bayesian Deep Learning Methods for Semantic Segmentationcited by 143
  5. 2019Detecting the Unexpected via Image Resynthesiscited by 137
  6. 2020Entropy Maximization and Meta Classification for Out-Of-Distribution Detection in Semantic Segmentationcited by 133
  7. 2022BEVSegFormer: Bird's Eye View Semantic Segmentation From Arbitrary Camera Rigscited by 123
  8. 2020Synthesize then Compare: Detecting Failures and Anomalies for Semantic Segmentationcited by 96
  9. 2023COTR: Compact Occupancy TRansformer for Vision-based 3D Occupancy Predictioncited by 96
  10. 2024SparseOcc: Rethinking Sparse Latent Representation for Vision-Based Semantic Occupancy Predictioncited by 96
  11. 2023Symphonize 3D Semantic Scene Completion with Contextual Instance Queriescited by 90
  12. 2023NDC-Scene: Boost Monocular 3D Semantic Scene Completion in Normalized Device Coordinates Spacecited by 90
  13. 2024GaussianFormer-2: Probabilistic Gaussian Superposition for Efficient 3D Occupancy Predictioncited by 86
  14. 2021Pixel-wise Energy-biased Abstention Learning for Anomaly Segmentation on Complex Urban Driving Scenescited by 83
  15. 2021MobileStereoNet: Towards Lightweight Deep Networks for Stereo Matchingcited by 82
  16. 2021Standardized Max Logits: A Simple yet Effective Approach for Identifying Unexpected Road Obstacles in Urban-Scene Segmentationcited by 77
  17. 2023SSCBench: A Large-Scale 3D Semantic Scene Completion Benchmark for Autonomous Drivingcited by 75
  18. 20253D and 4D World Modeling: A Surveycited by 67
  19. 2025WOD-E2E: Waymo Open Dataset for End-to-End Driving in Challenging Long-tail Scenarioscited by 62
  20. 2024DOME: Taming Diffusion Model into High-Fidelity Controllable Occupancy World Modelcited by 61
  21. 2024Tokenize the World into Object-level Knowledge to Address Long-tail Events in Autonomous Drivingcited by 56
  22. 2022Residual Pattern Learning for Pixel-wise Out-of-Distribution Detection in Semantic Segmentationcited by 55
  23. 2022Entropy-based Active Learning for Object Detection with Progressive Diversity Constraintcited by 53
  24. 2024POP-3D: Open-Vocabulary 3D Occupancy Prediction from Imagescited by 52
  25. 2021Deep Metric Learning for Open World Semantic Segmentationcited by 51
  26. 2024Context and Geometry Aware Voxel Transformer for Semantic Scene Completioncited by 50
  27. 2021Energy-Based Open-World Uncertainty Modeling for Confidence Calibrationcited by 46
  28. 2024Not All Voxels Are Equal: Hardness-Aware Semantic Scene Completion with Self-Distillationcited by 43
  29. 2023On Advantages of Mask-level Recognition for Outlier-aware Segmentationcited by 42
  30. 2025RAP: 3D Rasterization Augmented End-to-End Planningcited by 41
  31. 2022Towards Long-Tailed 3D Detectioncited by 40
  32. 2023Camera-based 3D Semantic Scene Completion with Sparse Guidance Networkcited by 40
  33. 2024Outlier detection by ensembling uncertainty with negative objectnesscited by 34
  34. 2024OccMamba: Semantic Occupancy Prediction with State Space Modelscited by 30
  35. 2025Semi-Supervised Vision-Centric 3D Occupancy World Model for Autonomous Drivingcited by 29
  36. 2022Augmenting Softmax Information for Selective Classification with Out-of-Distribution Datacited by 27
  37. 2023OVO: Open-Vocabulary Occupancycited by 24
  38. 2025RayFronts: Open-Set Semantic Ray Frontiers for Online Scene Understanding and Explorationcited by 23

50 most cited of 89 All 89 (JSON)

Built on by

Not cited yet.