Medical image analysis using deep learning: Recent advances, applications, challenges and future directions

Abstract

Deep learning is now causing a paradigm change in medical image analysis. This technology has lately gained the attention of the medical imaging community, prompting the organization of a dedicated conference on "Deep Learning based Medical Imaging". This survey of the literature looks at recent developments in the field and provides an insightful analysis of the key issues facing the field. This review report organizes the evaluated literature into sub-categories based on the underlying pattern recognition tasks and taxonomy based on the anatomy of humans. Without assuming a prior knowledge of deep learning, this literature review makes a significant contribution to educating non-experts in the medical community about the basic concepts of deep learning. In this literature review, the advancements of deep learning (DL) in medical image analysis are investigated from a distinct computer vision and machine learning viewpoint. This allows us to identify the "lack of correctly labeled huge scale data sets" as the primary problem in this study direction. The study also focuses on the sister research domains such as pattern recognition, computer vision, and machine learning. Also mentioned were approaches to finding plausible future scenarios in which the medical imaging community may fully utilize deep learning.