Out-of-Distribution Detection in LiDAR Semantic Segmentation Using Epistemic Uncertainty from Hierarchical GMMs

2025-10-08

AI Summary

To address out-of-distribution (OOD) object misclassification in LiDAR point cloud semantic segmentation, this paper proposes an unsupervised OOD detection method requiring no auxiliary data. The method operates solely on a single forward pass of a pre-trained segmentation network, without model ensembling or multi-stage training. Its core contribution is the first construction of a hierarchical Gaussian Mixture Model (GMM) in deep feature space, coupled with Bayesian inference to explicitly disentangle epistemic uncertainty (model ignorance) from aleatoric uncertainty (data randomness)—thereby overcoming the fundamental limitation of entropy-based methods that conflate these two uncertainty sources. Evaluated on SemanticKITTI, the approach achieves substantial improvements: +18% AUROC, +22% AUPRC, and a reduction in FPR95 from 76% to 40%, significantly outperforming state-of-the-art prediction-entropy baselines.

Technology Category

• Machine Learning: Calibration & Uncertainty Quantification
• Reasoning under Uncertainty: Uncertainty Representations
• Computer Vision: Segmentation

Application Category

• Semantics and Knowledge: Methods to enhance, augment, integrate or synergize semantic models such as knowledge graphs and LLMs
• Graph Algorithms and Modeling for the Web: Foundation models and LLMs for Web-related graphs
• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking

Abstract

In addition to accurate scene understanding through precise semantic segmentation of LiDAR point clouds, detecting out-of-distribution (OOD) objects, instances not encountered during training, is essential to prevent the incorrect assignment of unknown objects to known classes. While supervised OOD detection methods depend on auxiliary OOD datasets, unsupervised methods avoid this requirement but typically rely on predictive entropy, the entropy of the predictive distribution obtained by averaging over an ensemble or multiple posterior weight samples. However, these methods often conflate epistemic (model) and aleatoric (data) uncertainties, misclassifying ambiguous in distribution regions as OOD. To address this issue, we present an unsupervised OOD detection approach that employs epistemic uncertainty derived from hierarchical Bayesian modeling of Gaussian Mixture Model (GMM) parameters in the feature space of a deep neural network. Without requiring auxiliary data or additional training stages, our approach outperforms existing uncertainty-based methods on the SemanticKITTI dataset, achieving an 18% improvement in AUROC, 22% increase in AUPRC, and 36% reduction in FPR95 (from 76% to 40%), compared to the predictive entropy approach used in prior works.

Problem

• Detecting unknown objects in LiDAR semantic segmentation
• Separating epistemic and aleatoric uncertainties for OOD detection
• Improving unsupervised OOD detection without auxiliary datasets

Innovation

• Uses hierarchical GMMs for epistemic uncertainty estimation
• Extracts uncertainty from deep network feature space
• Eliminates need for auxiliary data or retraining