Chapter 2 - Deep learning basics

Abstract

We review several basic types of deep neural networks, including convolutional neural networks, recurrent neural networks, deep image-to-image networks, and deep generative networks. These networks are the core building blocks that we may use to develop deep learning based solutions to medical image computing. This chapter introduces the basic idea behind each network and also highlights some basic structures or models that are relevant to the topics of this book. In later chapters, we will demonstrate how these networks can be applied to address specific tasks and problems in medical image computing.