Feb 23, 2017
Green Hall, Room 0120
Faculty candidate, Yuejie Chi, PhD Assistant Professor Department of SBS-Biomedical Informatics and Electrical & Computer Engineering at The Ohio State University, will present.
Abstract: Source localization is an important problem across engineering and applied science, where the goal is to invert an acquired image for the underlying source signal that produced it. In this talk, driven by novel applications in imaging and sensor technologies, we consider source localization when conventional approaches become highly suboptimal or no longer apply, due to the existence of missing data, outliers, interference or lack of calibration in data acquisition. We will illustrate geometric data representations that exploit sparsity and physically-meaningful constraints, show how they lead to provably efficient and robust algorithms for source localization using the highly versatile framework of convex optimization, and demonstrate applications of some of these results in super-resolution microscopy imaging.