Improving Batch Normalization with TTA for Robust Object Detection in Self-Driving

TL;DR

This paper introduces two novel methods to enhance Batch Normalization with test-time adaptation for robust object detection in autonomous driving, addressing domain shifts caused by weather and sensor failures.

Contribution

It proposes a LearnableBN layer with GSEM and a semantic-consistency dual-stage adaptation strategy to improve TTA stability and performance in autonomous driving scenarios.

Findings

Achieved up to 8% performance improvement across various corruptions.

Demonstrated robustness of the proposed methods on NuScenes-C dataset.

Enhanced model stability during test-time adaptation.

Abstract

In current open real-world autonomous driving scenarios, challenges such as sensor failure and extreme weather conditions hinder the generalization of most autonomous driving perception models to these unseen domain due to the domain shifts between the test and training data. As the parameter scale of autonomous driving perception models grows, traditional test-time adaptation (TTA) methods become unstable and often degrade model performance in most scenarios. To address these challenges, this paper proposes two new robust methods to improve the Batch Normalization with TTA for object detection in autonomous driving: (1) We introduce a LearnableBN layer based on Generalized-search Entropy Minimization (GSEM) method. Specifically, we modify the traditional BN layer by incorporating auxiliary learnable parameters, which enables the BN layer to dynamically update the statistics according to the different input data. (2) We propose a new semantic-consistency based dual-stage-adaptation strategy, which encourages the model to iteratively search for the optimal solution and eliminates unstable samples during the adaptation process. Extensive experiments on the NuScenes-C dataset shows that our method achieves a maximum improvement of about 8% using BEVFormer as the baseline model across six corruption types and three levels of severity. We will make our source code available soon.