An adaptive Wiener filter for PET based on a novel technique for spectral estimation

Author

Sun, Winston Y. ; Derenzo, Stephen E. ; Budinger, Thomas F.

Author_Institution

California Univ., Berkeley, CA, USA

fYear

1991

fDate

2-9 Nov. 1991

Firstpage

1933

Abstract

An adaptive Wiener filter has been developed for the convolution backprojection (CBP) reconstruction of positron emission tomograph (PET) images. An approach is proposed for applying minimum mean square error (MMSE) producing filters in the image space by estimating the power spectra of the signal and noise in the reprojections of the reconstructed images. Backprojection-algorithm-dependent colorization of the white input noise process and angle-dependent signal are accounted for in the proposed algorithm to compute the Wiener filter. Reconstruction results using the proposed algorithm are compared with the manually selected Butterworth-type filters on data obtained with the Donner 600-crystal tomograph. Simulation results using the uniform circular disk phantom of estimating the noise and signal power spectra in the reconstructed images produced reasonable estimates of the Wiener filter.<>

Keywords

adaptive filters; computerised tomography; image reconstruction; radioisotope scanning and imaging; Butterworth-type filters; Donner 600-crystal tomograph; PET; adaptive Wiener filter; angle-dependent signal; backprojection-algorithm-dependent colorization; convolution backprojection reconstruction; medical diagnostic imaging; minimum mean square error; nuclear medicine; positron emission tomograph images; uniform circular disk phantom; white input noise process; Colored noise; Computational modeling; Convolution; Image reconstruction; Mean square error methods; Positron emission tomography; Radioactive decay; Signal processing; White noise; Wiener filter;

fLanguage

English

Publisher

ieee

Conference_Titel

Nuclear Science Symposium and Medical Imaging Conference, 1991., Conference Record of the 1991 IEEE

Conference_Location

Santa Fe, NM, USA

Print_ISBN

0-7803-0513-2

Type

conf

DOI

10.1109/NSSMIC.1991.259254

Filename

259254