Tomographic image reconstruction for systems with partially-known blur

Author

Brankov, Jovan G. ; Djordjevic, Jaksa ; Galatsanos, Nikolas P. ; Wernick, Miles N.

Author_Institution

Dept. of Electr. & Comput. Eng., Illinois Inst. of Technol., Chicago, IL, USA

Volume

3

fYear

1999

fDate

1999

Firstpage

881

Abstract

In tomographic image reconstruction, it is usually assumed that the system matrix is known exactly, although this is not usually the case in practice. We investigate the potential benefit of modeling the system matrix as the sum of a known part and an unknown random error. Using some of simplifying assumptions, we develop a penalized weighted least squares (PWLS) reconstruction algorithm for this problem. Our experiments indicate that this approach can, indeed lead to significant improvements in the reconstructed image, both visually and quantitatively

Keywords

computerised tomography; conjugate gradient methods; discrete Fourier transforms; image reconstruction; least squares approximations; conjugate gradient algorithm; discrete Fourier transform; maximum a posteriori; optical tomography; partially-known blur; penalized weighted least squares; positron emission tomography; random error; reconstruction algorithm; sinogram restoration; system matrix; tomographic image reconstruction; total least squares; Cost function; Covariance matrix; Discrete Fourier transforms; Image reconstruction; Image restoration; Least squares approximation; Least squares methods; Optical imaging; Positron emission tomography; Reconstruction algorithms;

fLanguage

English

Publisher

ieee

Conference_Titel

Image Processing, 1999. ICIP 99. Proceedings. 1999 International Conference on

Conference_Location

Kobe

Print_ISBN

0-7803-5467-2

Type

conf

DOI

10.1109/ICIP.1999.817276

Filename

817276