A Study of Using Deep Learning with Medical Images: Starting from Fundamental Artificial Neural Networks to Generative Models

Abstract

Recently, deep learning has shown significant progress and is advancing quickly in various automated applications with minimal errors. One such application is in medical image analysis for disease detection, where deep learning has demonstrated high accuracy and precision due to its automatic feature representations. This article presents an overview of essential deep learning concepts related to the generation of medical images. It offers brief summaries of research utilizing some of the most advanced models from recent years applied to medical images of various injured body areas or organs affected by diseases (e.g., brain tumors and COVID-19 lung pneumonia). The objective of this study is to provide a comprehensive summary of artificial neural networks (NNs) and deep generative models in medical imaging to encourage more groups and authors unfamiliar with deep learning to consider its use in medical research. Furthermore, it presents a compilation of commonly used public medical datasets containing magnetic resonance (MR) images, computed tomography (CT) scans, and standard images.