Stochastic Resonance Improves the Detection of Low Contrast Images in Deep Learning Models

TL;DR

This paper demonstrates that adding controlled noise to low-contrast images can enhance recognition performance in rate-based neural networks, revealing stochastic resonance as a useful phenomenon in image classification.

Contribution

It introduces the concept of stochastic resonance in rate-based neural networks and shows how noise can improve detection of low-contrast images.

Findings

Noise improves classification accuracy on low-contrast images

Stochastic resonance observed in rate-based neural networks

Potential for noise-based enhancement in neural network applications

Abstract

Stochastic resonance describes the utility of noise in improving the detectability of weak signals in certain types of systems. It has been observed widely in natural and engineered settings, but its utility in image classification with rate-based neural networks has not been studied extensively. In this analysis a simple LSTM recurrent neural network is trained for digit recognition and classification. During the test phase, image contrast is reduced to a point where the model fails to recognize the presence of a stimulus. Controlled noise is added to partially recover classification performance. The results indicate the presence of stochastic resonance in rate-based recurrent neural networks.