Query2Uncertainty: Robust Uncertainty Quantification and Calibration for 3D Object Detection under Distribution Shift

2026-05-06

AI Summary

This work addresses the challenge of unreliable uncertainty estimation and poor calibration in existing 3D object detectors under distribution shifts. The authors propose a density-aware calibration method that leverages, for the first time, the compactness, positional awareness, and class-sensitive characteristics of object queries in DETR-style detectors to jointly calibrate classification confidence and bounding box regression uncertainty through feature density. By integrating object-query-based density estimation with a post-hoc calibration module, the approach is applicable to both multi-view camera and LiDAR-based 3D detection architectures. Experiments demonstrate that the method significantly outperforms current post-hoc calibration techniques under both in-distribution and distribution-shift settings, effectively enhancing uncertainty reliability and overall model calibration performance.

Abstract

Reliable uncertainty estimation for 3D object detection is critical for deploying safe autonomous systems, yet modern detectors remain poorly calibrated, especially under distribution shifts. Although post-hoc calibration methods address this issue and provide improved calibration for in-distribution tests, they fail to adapt in distribution-shifted scenarios. In this work, we address this issue and introduce a density-aware calibration method that couples post-hoc calibrators with the feature density of latent object queries from DETR-style 3D object detectors. These queries form a compact, location and class-aware feature, ideal for density estimation, allowing our approach to adjust model confidences in distribution-shift scenarios. By fitting a density estimator on these query features, our approach jointly recalibrates both classification and bounding box regression uncertainties. On both a multi-view camera and LiDAR-based detector, our approach consistently outperforms standard post-hoc methods in both in-distribution and distribution-shifted scenarios. Code available https://tillbeemelmanns.github.io/query2uncertainty/ .

Problem

• uncertainty quantification
• 3D object detection
• distribution shift
• model calibration
• autonomous systems

Innovation

• uncertainty quantification
• distribution shift
• 3D object detection
• density estimation
• calibration