ABSTRACT
Deep learning algorithms have made considerable advancements in automated abnormality identification in radiological images, paving the way for their prospective use in computer-aided diagnostic systems. This chapter reviews recent advances in the development of powered deep learning computer vision systems for medical applications, concentrating on medical imaging, medical video, and clinical implementation. We begin by outlining a decade of progress in deep neural networks and the visual tasks they enable in the healthcare setting. Following that, we look at a variety of medical imaging applications that stand to benefit, such as cardiology, pneumology, neurology, and ophthalmology, and suggest new avenues for further study. Finally, we discuss the issues and challenges that must be overcome before these technologies can be employed in real-world clinical trials.
