Neural Distribution Prior for LiDAR Out-of-Distribution Detection

2026-04-10

AI Summary

Existing LiDAR perception models struggle to effectively detect out-of-distribution (OOD) objects in open-world settings, primarily due to severe class imbalance and the erroneous assumption of uniform class distribution. To address this, this work proposes the Neural Distribution Prior (NDP) framework, which leverages an attention mechanism to model the structural properties of logit distributions and adaptively corrects class-specific confidence biases. Additionally, it introduces an input-level OOD sample synthesis strategy based on Perlin noise, enhancing training robustness without requiring external data. The NDP framework seamlessly integrates with various OOD scoring mechanisms and achieves substantial performance gains on both SemanticKITTI and STU benchmarks. Notably, it attains a point-level average precision of 61.31% on the STU test set—more than tenfold improvement over the previous state-of-the-art method.

Technology Category

• Machine Learning: Calibration & Uncertainty Quantification
• Computer Vision: Diffusion Models for Vision
• Constraint Satisfaction and Optimization: Distributed CSP/Optimization

Application Category

• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking
• 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

Abstract

LiDAR-based perception is critical for autonomous driving due to its robustness to poor lighting and visibility conditions. Yet, current models operate under the closed-set assumption and often fail to recognize unexpected out-of-distribution (OOD) objects in the open world. Existing OOD scoring functions exhibit limited performance because they ignore the pronounced class imbalance inherent in LiDAR OOD detection and assume a uniform class distribution. To address this limitation, we propose the Neural Distribution Prior (NDP), a framework that models the distributional structure of network predictions and adaptively reweights OOD scores based on alignment with a learned distribution prior. NDP dynamically captures the logit distribution patterns of training data and corrects class-dependent confidence bias through an attention-based module. We further introduce a Perlin noise-based OOD synthesis strategy that generates diverse auxiliary OOD samples from input scans, enabling robust OOD training without external datasets. Extensive experiments on the SemanticKITTI and STU benchmarks demonstrate that NDP substantially improves OOD detection performance, achieving a point-level AP of 61.31\% on the STU test set, which is more than 10$\times$ higher than the previous best result. Our framework is compatible with various existing OOD scoring formulations, providing an effective solution for open-world LiDAR perception.

Problem

• LiDAR
• out-of-distribution detection
• class imbalance
• open-world perception
• OOD scoring

Innovation

• Neural Distribution Prior
• Out-of-Distribution Detection
• LiDAR Perception
• Class Imbalance
• Perlin Noise Synthesis