Solving phase retrieval via graph projection splitting

DOI10.1088/1361-6420/AB79FAzbMATH Open1471.90118arXiv1910.08714OpenAlexW3007493423MaRDI QIDQ5000597FDOQ5000597

Authors: Ji Li, Hongkai Zhao 

Publication date: 14 July 2021

Published in: Inverse Problems (Search for Journal in Brave)

Abstract: Phase retrieval with prior information can be cast as a nonsmooth and nonconvex optimization problem. We solve the problem by graph projection splitting (GPS), where the two proximity subproblems and the graph projection step can be solved efficiently. With slight modification, we also propose a robust graph projection splitting (RGPS) method to stabilize the iteration for noisy measurements. Contrary to intuition, RGPS outperforms GPS with fewer iterations to locate a satisfying solution even for noiseless case. Based on the connection between GPS and Douglas-Rachford iteration, under mild conditions on the sampling vectors, we analyze the fixed point sets and provide the local convergence of GPS and RGPS applied to noiseless phase retrieval without prior information. For noisy case, we provide the error bound of the reconstruction. Compared to other existing methods, thanks for the splitting approach, GPS and RGPS can efficiently solve phase retrieval with prior information regularization for general sampling vectors which are not necessarily isometric. For Gaussian phase retrieval, compared to existing gradient flow approaches, numerical results show that GPS and RGPS are much less sensitive to the initialization. Thus they markedly improve the phase transition in noiseless case and reconstruction in the presence of noise respectively. GPS shows sharpest phase transition among existing methods including RGPS, while it needs more iterations than RGPS when the number of measurement is large enough. RGPS outperforms GPS in terms of stability for noisy measurements. When applying RGPS to more general non-Gaussian measurements with prior information, such as support, sparsity and TV minimization, RGPS either outperforms state-of-the-art solvers or can be combined with state-of-the-art solvers to improve their reconstruction quality.

Full work available at URL: https://arxiv.org/abs/1910.08714

Recommendations

• Phase retrieval with sparse phase constraint
• Phase retrieval with background information
• Phase retrieval via Wirtinger flow: theory and algorithms
• Variational phase retrieval with globally convergent preconditioned proximal algorithm
• Phase retrieval for sub-Gaussian measurements

zbMATH Keywords

local convergencephase retrievalDouglas-Rachford iterationgraph projection splitting

Mathematics Subject Classification ID

Applications of mathematical programming (90C90) Nonconvex programming, global optimization (90C26)

Cites Work

• Phaselift: exact and stable signal recovery from magnitude measurements via convex programming
• Phase retrieval via Wirtinger flow: theory and algorithms
• Alternating direction methods for classical and ptychographic phase retrieval
• Relaxed averaged alternating reflections for diffraction imaging
• Phase recovery, MaxCut and complex semidefinite programming
• Phase retrieval from coded diffraction patterns
• Phase retrieval via matrix completion
• Solving Systems of Random Quadratic Equations via Truncated Amplitude Flow
• The reconstruction of a multidimensional sequence from the phase or magnitude of its Fourier transform
• A geometric analysis of phase retrieval
• Solving Random Quadratic Systems of Equations Is Nearly as Easy as Solving Linear Systems
• Fourier phase retrieval with a single mask by Douglas-Rachford algorithms
• Phase Retrieval Using Alternating Minimization
• On global convergence of gradient descent algorithms for generalized phase retrieval problem
• On relaxed averaged alternating reflections (RAAR) algorithm for phase retrieval with structured illumination
• Phase Retrieval by Alternating Minimization With Random Initialization

Cited In (3)

• Solving blind ptychography effectively via linearized alternating direction method of multipliers
• Scalable incremental nonconvex optimization approach for phase retrieval
• Relax-and-split method for nonconvex inverse problems

Uses Software

• PhasePack
• PhaseLift
• Wirtinger Flow

This page was built for publication: Solving phase retrieval via graph projection splitting

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q5000597)