Dead Pixel Test Using Effective Receptive Field

08/31/2021
by   Bum Jun Kim, et al.
0

Deep neural networks have been used in various fields, but their internal behavior is not well known. In this study, we discuss two counterintuitive behaviors of convolutional neural networks (CNNs). First, we evaluated the size of the receptive field. Previous studies have attempted to increase or control the size of the receptive field. However, we observed that the size of the receptive field does not describe the classification accuracy. The size of the receptive field would be inappropriate for representing superiority in performance because it reflects only depth or kernel size and does not reflect other factors such as width or cardinality. Second, using the effective receptive field, we examined the pixels contributing to the output. Intuitively, each pixel is expected to equally contribute to the final output. However, we found that there exist pixels in a partially dead state with little contribution to the output. We reveal that the reason for this lies in the architecture of CNN and discuss solutions to reduce the phenomenon. Interestingly, for general classification tasks, the existence of dead pixels improves the training of CNNs. However, in a task that captures small perturbation, dead pixels degrade the performance. Therefore, the existence of these dead pixels should be understood and considered in practical applications of CNN.

READ FULL TEXT

page 4

page 6

research
01/15/2017

Understanding the Effective Receptive Field in Deep Convolutional Neural Networks

We study characteristics of receptive fields of units in deep convolutio...
research
11/17/2022

Convolutional neural networks for medical image segmentation

In this article, we look into some essential aspects of convolutional ne...
research
09/13/2022

Analyzing the Impact of Varied Window Hyper-parameters on Deep CNN for sEMG based Motion Intent Classification

The use of deep neural networks in electromyogram (EMG) based prostheses...
research
11/26/2022

Receptive Field Refinement for Convolutional Neural Networks Reliably Improves Predictive Performance

Minimal changes to neural architectures (e.g. changing a single hyperpar...
research
05/31/2022

SAR Despeckling Using Overcomplete Convolutional Networks

Synthetic Aperture Radar (SAR) despeckling is an important problem in re...
research
05/29/2019

Learning the Non-linearity in Convolutional Neural Networks

We propose the introduction of nonlinear operation into the feature gene...
research
07/30/2019

4X4 Census Transform

This paper proposes a 4X4 Census Transform (4X4CT) to encourage further ...

Please sign up or login with your details

Forgot password? Click here to reset