The performance of the amplitude-based model for complex phase retrieval

DOI10.1093/IMAIAI/IAAD053arXiv2204.05492OpenAlexW4391057899WikidataQ129640118 ScholiaQ129640118MaRDI QIDQ6396263FDOQ6396263

Authors: Yu Xia, Zhiqiang Xu 

Publication date: 11 April 2022

Abstract: The aim of this paper is to study the performance of the amplitude-based model $widehatxin{margmin}_{xinmathbb{C}^d}su{m}_{j=1}^{m}left(|langle{a}_j,xangle|-b_{j}ight{)}^2$, where $b_j=|langle{a}_j,x_{0}angle|+et{a}_j$ and $x_{0}in{mathbbC}^d$ is a target signal. The model is raised in phase retrieval and one has developed many efficient algorithms to solve it. However, there are very few results about the estimation performance in complex case. We show that $mi{n}_{hetain[0,2pi)}|widehatx-exp(iheta)cdot{x}_0{|}_{2}lesssimfrac|eta{|}_{2}sqrtm$ holds with high probability provided the measurement vectors $a_{j}in{mathbbC}^d,$ $j=1,ldots,m,$ are complex Gaussian random vectors and $mgtrsimd$. Here $eta=(et{a}_1,ldots,et{a}_m)inmathbb{R}^m$ is the noise vector without any assumption on the distribution. Furthermore, we prove that the reconstruction error is sharp. For the case where the target signal $x_{0}inmathbb{C}^d$ is sparse, we establish a similar result for the nonlinear constrained LASSO. This paper presents the first theoretical guarantee on quadratic models for complex phase retrieval. To accomplish this, we leverage a strong version of restricted isometry property for low-rank matrices.

Full work available at URL: https://doi.org/10.1093/imaiai/iaad053

Mathematics Subject Classification ID

Signal theory (characterization, reconstruction, filtering, etc.) (94A12)

This page was built for publication: The performance of the amplitude-based model for complex phase retrieval

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q6396263)