학술논문
Image quality for diffraction tomography, holographic backpropagation, and regularized sampling with noisy data.
Document Type
Article
Author
Source
Subject
*IMAGE quality analysis
*TOMOGRAPHY
*OPTICAL diffraction
*BACK propagation
*HOLOGRAPHY
*IMAGE reconstruction
*STATISTICAL sampling
*GEOMETRY
*
*
*
*
*
*
*
Language
ISSN
1741-5977
Abstract
We consider the behaviour of inverse image reconstruction quality as a function of forward data SNR. A canonical scatterer geometry is used to obtain an analytical image contrast ratio estimate for Colton-Kirsch regularized sampling. For holographic backpropagation tomography, an estimate is developed for the SNR at which the image contrast ratio transitions from noise-dominated to the noise-free contrast ratio limit. Numerical results confirm the theoretical analyses and extend the results to more complex geometries. Holographic backpropagation exhibits the greatest noise immunity in the sense that the reconstructed image contrast ratio reaches an acceptable threshold at a lower SNR than is the case for standard diffraction tomography or the regularized sampling method. [ABSTRACT FROM AUTHOR]