Video Description

Dr. Emrah Bostan presents a tutorial on the Fundamentals of Algorithms for Computational Imaging. In this talk, we consider practical algorithms for inverse problems in imaging. Adopting a suitable language, the concept is to “invert” the data acquisition process that relates an unknown image to observable measurements. By doing so, one obtains the spatial distribution of physical parameters. Since we retrieve information about entities—via solving inverse problems—that we are unable to observe directly, the topic is among the most intensively studied mathematical problems in the field. To have an expanded perspective, we shall be placed at the interplay of the deterministic and the stochastic approaches. Starting from linear image reconstruction algorithms, we shall review recent techniques that are nonlinear. We will present examples of such algorithms used in practical cases such as phase retrieval and lensless imaging.