TY - JOUR
T1 - Development and evaluation of MRI based bayesian image reconstruction methods for PET
AU - Wang, Chao Hsing
AU - Chen, Jyh-Cheng
AU - Kao, Chien Min
AU - Liu, Ren-Shyan
N1 - Funding Information:
This work was supported in part by VTY Joint Research Program, Tsou's Foundation under Grant VTY90-P2-14. The first author was also supported in part by NRPGM under Grant NSC91-3112-P-075-001-Y. The authors wish to thank C.M. Kao for providing invaluable information. We thank H.H.S. Lu, and I.T. Hsiao for their invaluable discussions. The authors also appreciate the help of T.B. Chen, K.L. Chou, K.Y. Tu, R.M. Hwang, and H.S. Chueh in preparing the experiments. We thank Eric Jao for his help in editing.
PY - 2003/10/19
Y1 - 2003/10/19
N2 - MRI Based Bayesian Image Reconstruction Methods for positron emission tomography reconstruction with the aim at improving image quality was developed. The line site map from MRI as a priori is made up of either a modified Markov random field or Canny edge detector with the Gaussian smoothing filter. It is used in maximum a posteriori (MAP) algorithm by a weighted line site method. We evaluate and compare the performance of these reconstruction methods. The results show that the Bayesian methods produce reconstructed images with less noise and better spatial resolution than those produced by the maximum likelihood-expectation maximization method.
AB - MRI Based Bayesian Image Reconstruction Methods for positron emission tomography reconstruction with the aim at improving image quality was developed. The line site map from MRI as a priori is made up of either a modified Markov random field or Canny edge detector with the Gaussian smoothing filter. It is used in maximum a posteriori (MAP) algorithm by a weighted line site method. We evaluate and compare the performance of these reconstruction methods. The results show that the Bayesian methods produce reconstructed images with less noise and better spatial resolution than those produced by the maximum likelihood-expectation maximization method.
UR - http://www.scopus.com/inward/record.url?scp=11844296722&partnerID=8YFLogxK
U2 - 10.1109/NSSMIC.2003.1352422
DO - 10.1109/NSSMIC.2003.1352422
M3 - Conference article
AN - SCOPUS:11844296722
SN - 1095-7863
VL - 4
SP - 2599
EP - 2603
JO - IEEE Nuclear Science Symposium Conference Record
JF - IEEE Nuclear Science Symposium Conference Record
M1 - M10-220
T2 - 2003 IEEE Nuclear Science Symposium Conference Record - Nuclear Science Symposium, Medical Imaging Conference
Y2 - 19 October 2003 through 25 October 2003
ER -