Quadratic mutual information regularization in real-time deep CNN models

TL;DR

This paper introduces a new regularization technique based on Quadratic Mutual Information to enhance the generalization of lightweight deep CNNs designed for real-time high-resolution video processing on resource-limited devices.

Contribution

The paper proposes a novel regularization method inspired by Quadratic Mutual Information and demonstrates its effectiveness in improving lightweight CNN models for real-time applications.

Findings

Improved generalization in binary classification tasks

Effective real-time performance on resource-constrained devices

Regularizer enhances model robustness

Abstract

In this paper, regularized lightweight deep convolutional neural network models, capable of effectively operating in real-time on devices with restricted computational power for high-resolution video input are proposed. Furthermore, a novel regularization method motivated by the Quadratic Mutual Information, in order to improve the generalization ability of the utilized models is proposed. Extensive experiments on various binary classification problems involved in autonomous systems are performed, indicating the effectiveness of the proposed models as well as of the proposed regularizer.