DeepAI AI Chat
Log In Sign Up

Improved Robustness of Vision Transformer via PreLayerNorm in Patch Embedding

11/16/2021
by   Bum Jun Kim, et al.
POSTECH / 포항공과대학교
0

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.

READ FULL TEXT
12/21/2021

MPViT: Multi-Path Vision Transformer for Dense Prediction

Dense computer vision tasks such as object detection and segmentation re...
04/26/2022

Deeper Insights into ViTs Robustness towards Common Corruptions

Recent literature have shown design strategies from Convolutions Neural ...
06/05/2021

Patch Slimming for Efficient Vision Transformers

This paper studies the efficiency problem for visual transformers by exc...
03/26/2023

Sector Patch Embedding: An Embedding Module Conforming to The Distortion Pattern of Fisheye Image

Fisheye cameras suffer from image distortion while having a large field ...
10/15/2021

Understanding and Improving Robustness of Vision Transformers through Patch-based Negative Augmentation

We investigate the robustness of vision transformers (ViTs) through the ...
06/30/2020

BitMix: Data Augmentation for Image Steganalysis

Convolutional neural networks (CNN) for image steganalysis demonstrate b...
07/31/2017

Analysis and Optimization of Convolutional Neural Network Architectures

Convolutional Neural Networks (CNNs) dominate various computer vision ta...