Recent mathematical advances in phase retrieval and computational imagingMS21

In many areas of science, one has access to magnitude only measurements. Phase retrieval is the problem of recovering a signal from such measurements. Due to its practical significance in imaging science, numerous heuristics have been developed to solve such problems. Novel theoretical developments and exciting new algorithms and applications have led to a renewed interest in phase retrieval. The proposed minisymposium focuses on recent theoretical results showing the success of nonconvex optimization algorithms as well as spectacular recent research efforts in imaging applications. A special emphasis is on theoretical and algorithmic developments that are tightly related to real-world applications.

Wed 06 June at 09:30 in Room I (Palazzina B - Building B floor 0)
Denoising with spherically uniform neural layers
Dustin Mixon ( Department of Mathematics, Ohio State University )
Fast Phase Retrieval from Windowed Fourier Measurements via Wigner Distribution Deconvolution + Angular Synchronization
Mark Iwen ( Department of Mathematics, Michigan State University )
Nonconvex Sparse Blind Deconvolution: Geometry and Efficient Methods
John Wright ( Department of Electrical Engineering, Columbia University )
Invariants for cryo-EM and multireference alignment
Tamir Bendory ( Princeton University )
Wed 06 June at 13:00 in Room I (Palazzina B - Building B floor 0)
Phase retrieval by alternating projections for random Gaussian measurements
Irene Waldspurger ( CEREMADE (Université Paris-Dauphine) )
Phase retrieval with structured measurements
Yonina Eldar ( Department of EE, Technion, Israel Institute of Technology, Haifa )
Phase Retrieval Without Small-Ball Probability Assumptions
Felix Krahmer ( Technical University of Munich, Department of Mathematics )
What are heuristic phase retrieval algorithms and why you should care
Veit Elser ( Department of Physics, Cornell University )
Tamir Bendory ( Princeton University )
Mahdi Soltanolkotabi ( University of Southern California )
image deblurring, image reconstruction, inverse problems, machine learning, nonlinear optimization