Effects of Degradations on Deep Neural Network Architectures

TL;DR

This paper systematically evaluates how six deep neural network architectures perform under various common image degradations, providing insights into their robustness for real-world applications.

Contribution

It offers an extensive comparative analysis of six popular CNN architectures' robustness to multiple image noise and degradation models.

Findings

ResNet-50 shows high robustness to Gaussian noise.

CapsuleNet performs best under salt-and-pepper noise.

Inception-v3 maintains stability across multiple degradation types.

Abstract

Deep convolutional neural networks (CNN) have massively influenced recent advances in large-scale image classification. More recently, a dynamic routing algorithm with capsules (groups of neurons) has shown state-of-the-art recognition performance. However, the behavior of such networks in the presence of a degrading signal (noise) is mostly unexplored. An analytical study on different network architectures toward noise robustness is essential for selecting the appropriate model in a specific application scenario. This paper presents an extensive performance analysis of six deep architectures for image classification on six most common image degradation models. In this study, we have compared VGG-16, VGG-19, ResNet-50, Inception-v3, MobileNet and CapsuleNet architectures on Gaussian white, Gaussian color, salt-and-pepper, Gaussian blur, motion blur and JPEG compression noise models.