Fast Phase Retrieval from Deterministic Measurements in Two Dimensions

Thursday, August 17, 2017 - 10:30am - 11:00am
Lind 305
Brian Preskitt (University of California, San Diego)
The phase retrieval problem has appeared in a multitude of applications for decades. While ad hoc solutions have existed since the early 1970s, recent developments have provided algorithms that offer promising theoretical guarantees under increasingly realistic assumptions. Motivated by ptychographic imaging, we generalize a recent result on phase retrieval of a one dimensional signal to recover a two dimensional image from phaseless measurements, using a tensor product formulation to extend the previous work.