Data-Driven Computational Optical Imaging

It is well known that neural networks including deep learning have been widely employed to solve the problems in recognition and classification. It was not until recently that people started to use them to solve imaging problems. In this lecture, I will first introduce the basic concept of deep learning, and then discuss how it can be adopted to solve computational imaging problems. In particular, I will talk about three use cases: computational ghost imaging, imaging through scattering media, and digital holography.