Enhanced Model Robustness to Input Corruptions by Per-corruption Adaptation of Normalization Statistics

TL;DR

This paper introduces PAN, a method that dynamically adapts normalization statistics based on identified corruption types to improve the robustness of vision models in challenging environments, outperforming existing solutions.

Contribution

The paper proposes a novel per-corruption adaptation technique for normalization statistics that enhances model robustness across multiple vision tasks and datasets.

Findings

PAN improves robustness on real-world corrupted datasets by 20-30%.

It outperforms baseline models in object recognition benchmarks.

Seamlessly integrates with existing convolutional models.

Abstract

Developing a reliable vision system is a fundamental challenge for robotic technologies (e.g., indoor service robots and outdoor autonomous robots) which can ensure reliable navigation even in challenging environments such as adverse weather conditions (e.g., fog, rain), poor lighting conditions (e.g., over/under exposure), or sensor degradation (e.g., blurring, noise), and can guarantee high performance in safety-critical functions. Current solutions proposed to improve model robustness usually rely on generic data augmentation techniques or employ costly test-time adaptation methods. In addition, most approaches focus on addressing a single vision task (typically, image recognition) utilising synthetic data. In this paper, we introduce Per-corruption Adaptation of Normalization statistics (PAN) to enhance the model robustness of vision systems. Our approach entails three key components: (i) a corruption type identification module, (ii) dynamic adjustment of normalization layer statistics based on identified corruption type, and (iii) real-time update of these statistics according to input data. PAN can integrate seamlessly with any convolutional model for enhanced accuracy in several robot vision tasks. In our experiments, PAN obtains robust performance improvement on challenging real-world corrupted image datasets (e.g., OpenLoris, ExDark, ACDC), where most of the current solutions tend to fail. Moreover, PAN outperforms the baseline models by 20-30% on synthetic benchmarks in object recognition tasks.