1 April 2016 Fast conjugate gradient algorithm extension for analyzer-based imaging reconstruction
Author Affiliations +
Abstract
This paper presents an extension of the classic Conjugate Gradient Algorithm. Motivated by the Analyzer-Based Imaging inverse problem, the novel method maximizes the Poisson regularized log-likelihood with a non-linear transformation of parameter faster than other solutions. The new approach takes advantage of the special properties of the Poisson log-likelihood to conjugate each ascend direction with respect all the previous directions taken by the algorithm. Our solution is compared with the general solution for non-quadratic unconstrained problems: the Polak- Ribiere formula. Both methods are applied to the ABI reconstruction problem.
© (2016) COPYRIGHT Society of Photo-Optical Instrumentation Engineers (SPIE). Downloading of the abstract is permitted for personal use only.
Oriol Caudevilla, Oriol Caudevilla, Jovan G. Brankov, Jovan G. Brankov, } "Fast conjugate gradient algorithm extension for analyzer-based imaging reconstruction", Proc. SPIE 9783, Medical Imaging 2016: Physics of Medical Imaging, 97834J (1 April 2016); doi: 10.1117/12.2217164; https://doi.org/10.1117/12.2217164
PROCEEDINGS
6 PAGES


SHARE
Back to Top