Abstract The emergence of modern imaging techniques and an unprecedented amount of computational power offers the opportunity to analyze high-dimensional medical imaging data in ways that previously were not possible. In this context, high-throughput data-driven approaches become important for discovering knowledge from large-scale medical imaging data. Since deep neural networks can learn simple concepts first and then build up more complex concepts in a layer-by-layer manner, deep learning techniques have resulted in many successful applications in the medical imaging area. In this chapter, we first briefly overview various deep learning models and their learning principles. Then, we demonstrate the hands-on experience of developing deep neural networks for medical image analysis with several typical applications such as feature representation learning, image segmentation, and image registration.