DOLPHIn -- Dictionary Learning for Phase Retrieval

Andreas Tillmann, Yonina Eldar, Julien Mairal
IEEE Transactions on Signal Processing

We propose a new algorithm to learn a dictionary for reconstructing and sparsely encoding signals from measurements without phase. Specifically, we consider the task of estimating a two-dimensional image from squared-magnitude measurements of a complex-valued linear transformation of the original image. Several recent phase retrieval algorithms exploit underlying sparsity of the unknown signal in order to improve recovery performance. In this work, we consider such a sparse signal prior in the context of phase retrieval, when the sparsifying dictionary is not known in advance. Our algorithm jointly reconstructs the unknown signal-possibly corrupted by noise-and learns a dictionary such that each patch of the estimated image can be sparsely represented. Numerical experiments demonstrate that our approach can obtain significantly better reconstructions for phase retrieval problems with noise than methods that cannot exploit such “hidden” sparsity. Moreover, on the theoretical side, we provide a convergence result for our method.

Accompanying software to be found below:
  • DOLPHIn - Matlab implementation of the dictionary learning method for 2D noisy (sparse) phase retrieval. (Version 1.10 of 07/25/2016; requires SPAMS package obtainable from http://spams-devel.gforge.inria.fr/)
Moreover, below you can also find
  • the test images from the paper
  • a document with lots of result tables of supplementary numerical experiments
» Show BibTeX

author = {A. M. Tillmann and Y. C. Eldar and J. Mairal},
title = {{DOLPHIn -- Dictionary Learning for Phase Retrieval}},
journal = {{IEEE Transactions on Signal Processing}},
volume = {64},
number = {24},
pages = {6485--6500},
year = {2016},
note = {DOI: 10.1109/TSP.2016.2607180}

Disclaimer Home Visual Computing institute RWTH Aachen University