Cost-Sensitive Uncertainty-Based Failure Recognition for Object Detection

TL;DR

This paper introduces a cost-sensitive, uncertainty-based framework for object detection that optimizes failure recognition rates within user-defined error budgets, significantly improving safety in autonomous driving scenarios.

Contribution

It proposes a novel cost-sensitive thresholding method that automates and optimizes failure detection based on uncertainty, tailored to specific error budgets.

Findings

Boosts failure recognition rate by 36-60%

Enhances safety in autonomous driving datasets

Utilizes localization aleatoric uncertainty and softmax entropy

Abstract

Object detectors in real-world applications often fail to detect objects due to varying factors such as weather conditions and noisy input. Therefore, a process that mitigates false detections is crucial for both safety and accuracy. While uncertainty-based thresholding shows promise, previous works demonstrate an imperfect correlation between uncertainty and detection errors. This hinders ideal thresholding, prompting us to further investigate the correlation and associated cost with different types of uncertainty. We therefore propose a cost-sensitive framework for object detection tailored to user-defined budgets on the two types of errors, missing and false detections. We derive minimum thresholding requirements to prevent performance degradation and define metrics to assess the applicability of uncertainty for failure recognition. Furthermore, we automate and optimize the thresholding process to maximize the failure recognition rate w.r.t. the specified budget. Evaluation on three autonomous driving datasets demonstrates that our approach significantly enhances safety, particularly in challenging scenarios. Leveraging localization aleatoric uncertainty and softmax-based entropy only, our method boosts the failure recognition rate by 36-60\% compared to conventional approaches. Code is available at https://mos-ks.github.io/publications.