Phase retrieval (PR) algorithms play a key role in many modern computational imagers. For instance, in ptychography and coherent inverse scattering, PR algorithms enable imaging past the diffraction limit and through multiple scattering materials. We study the tradeoffs associated with the panoply of PR algorithms, including the required number of measurements, computational complexity, and robustness to the measurement data distribution. To address the shortcomings of current methods, we introduce and test a new robust PR algorithm that is significantly faster.

This presentation is part of Minisymposium “MS60 - Computational and Compressive Imaging Technologies and Applications (3 parts)”
organized by: Robert Muise (Lockheed Martin) , Richard Baraniuk (Rice University) .

Authors:

Richard Baraniuk (Rice University)

Ashok Veeraraghavan (Rice University)

Christopher Metzler (Rice University)

Keywords:

image reconstruction, inverse problems, nonlinear optimization, phase retrieval, statistical inverse estimation methods