EBSCO Logo
Connecting you to content on EBSCOhost
Title

Phase‐Retrieval Algebraic Solution Based on Window Modulation.

Authors

Huang, Jianhui; Jin, Huiliang; Ye, Qian; Meng, Guoxiang

Abstract

Phase retrieval is to recover the object amplitude and phase from the diffraction patterns. Here, an algebraic method is presented to solve the phase retrieval problems of arbitrary 2D complex‐valued objects, without the requirements of numerous iterations, object prior knowledge, and precise experimental setups. The key step is to extract the far‐field phase from the diffraction patterns. It is demonstrated by rigorous theoretical derivation that three window‐shaped modulations are sufficient to establish a system of linear equations related to the far‐field phase, from which the far‐field phase is solved approximately with only algebraic calculations. Then, an efficient iteration scheme is designed to optimize the direct approximate solution. After a few iterations, the direct solution is improved quickly and very close to the exact solution. With the advantages of high speed and high accuracy to general phase retrieval problems, the proposed algebraic method may be helpful to the Fourier imaging systems in which interferometric techniques are not applicable. A general algebraic phase retrieval solution is found to reconstruct complex objects from diffraction patterns. Three window modulations are applied to establish a linear equation, from which the far‐field phase can be solved directly. Then with iteration optimization or object preprocessing, the initial approximate solution is improved highly and close to the exact solution.

Subjects

LIGHT modulators; LINEAR equations; ITERATIVE methods (Mathematics); APPROXIMATION theory; PROBLEM solving

Publication

Annalen der Physik, 2018, Vol 530, Issue 10, pN.PAG

ISSN

0003-3804

Publication type

Academic Journal

DOI

10.1002/andp.201800063

EBSCO Connect | Privacy policy | Terms of use | Copyright | Manage my cookies
Journals | Subjects | Sitemap
© 2025 EBSCO Industries, Inc. All rights reserved