Improved Robustness of Vision Transformer via PreLayerNorm in Patch Embedding

TL;DR

This paper enhances Vision Transformer robustness by introducing PreLayerNorm in patch embedding, addressing scale-invariance issues, and demonstrating improved performance across various image corruptions, especially contrast variations.

Contribution

The paper proposes a novel PreLayerNorm patch embedding method for ViT, improving its robustness to scale changes and contrast variations compared to standard ViT.

Findings

ViT with PreLayerNorm outperforms standard ViT in robustness tests.

PreLayerNorm mitigates performance degradation caused by contrast enhancement.

ViT with PreLayerNorm maintains higher accuracy across various image corruptions.

Abstract

Vision transformers (ViTs) have recently demonstrated state-of-the-art performance in a variety of vision tasks, replacing convolutional neural networks (CNNs). Meanwhile, since ViT has a different architecture than CNN, it may behave differently. To investigate the reliability of ViT, this paper studies the behavior and robustness of ViT. We compared the robustness of CNN and ViT by assuming various image corruptions that may appear in practical vision tasks. We confirmed that for most image transformations, ViT showed robustness comparable to CNN or more improved. However, for contrast enhancement, severe performance degradations were consistently observed in ViT. From a detailed analysis, we identified a potential problem: positional embedding in ViT's patch embedding could work improperly when the color scale changes. Here we claim the use of PreLayerNorm, a modified patch embedding structure to ensure scale-invariant behavior of ViT. ViT with PreLayerNorm showed improved robustness in various corruptions including contrast-varying environments.