Comprehensive Survey of Classical and Quantum Image Compression Approaches using Neural Networks

Abstract

The realm of digital signal processing places significant emphasis on image compression, with recent years seeing a surge of interest in neural network-based methods. This survey paper aims to offer a comprehensive review of cutting-edge techniques for image compression that utilize neural networks. The introduction of the study provides an overview of conventional picture compression techniques and their inherent drawbacks. The study then explores the idea of deep learning and how it might be used for picture compression. The survey outlines the various types of neural networks that are utilized for image compression, classifying them based on their compression objective and examining the various training approaches. Furthermore, the paper provides a comparative analysis of several state-of-the-art neural network-based image compression methods by drawing upon literature. This survey paper explores the potential of combining neural networks and quantum computing in the field of image compression, specifically through the use of quantum convolutional neural networks (QCNNs). By leveraging the parallel processing power of quantum computing and the ability of neural networks to recognize patterns in image data, QCNNs offer a promising approach for developing more efficient and effective compression algorithms that preserve image quality while reducing file size. Ultimately, the paper concludes with a summary of the survey, offering insights into the potential of neural network-based image compression for future research. Researchers interested in this area may utilize this survey as a reference.