Instance-Level Post Hoc Uncertainty Quantification in Object Detection

2026-06-03

AI Summary

In safety-critical applications such as autonomous driving, object detection models must provide instance-level bounding box uncertainty to ensure reliability. This work proposes MC-GLM, a Monte Carlo Generalized Linearization Model based on Laplace approximation, which enables efficient post-hoc uncertainty estimation without requiring model retraining. Notably, MC-GLM achieves constant sampling complexity for the first time, supports parallel computation, and avoids the drawbacks of multiple backpropagations or non-fully post-hoc procedures. Experiments on the nuScenes dataset using CenterPoint demonstrate that the resulting uncertainties are of high quality while maintaining computational efficiency and practical deployability.

Abstract

Object detection is a safety-critical component of autonomous driving. It is essential to quantify the uncertainty in bounding-box predictions for safety assurance. Post hoc uncertainty quantification without retraining aligns with real-world deployment requirements; therefore, we employ the Laplace approximation. Because instance-level uncertainty is needed, linearized inference methods that require multiple backpropagations are not time-efficient, and sampling-based methods are not fully post hoc. We propose Monte-Carlo generalized linearized model (MC-GLM), which provides instance-level and approximately post hoc uncertainty quantification. The number of samples required in the Monte Carlo step is constant and independent of the number of output instances, so it can be parallelized. Experiments on the nuScenes dataset with the CenterPoint detector validate the effectiveness of our method, and the resulting uncertainties exhibit good quality.

Problem

• uncertainty quantification
• object detection
• instance-level
• post hoc
• autonomous driving

Innovation

• Monte Carlo
• generalized linearized model
• instance-level uncertainty
• post hoc uncertainty quantification
• object detection