Title :
Globally convergent algorithms for maximum a posteriori transmission tomography
Author :
Lange, Kenneth ; Fessler, Jeffrey A.
Author_Institution :
Dept. of Biostat., Michigan Univ., Ann Arbor, MI, USA
fDate :
10/1/1995 12:00:00 AM
Abstract :
This paper reviews and compares three maximum likelihood algorithms for transmission tomography. One of these algorithms is the EM algorithm, one is based on a convexity argument devised by De Pierro (see IEEE Trans. Med. Imaging, vol.12, p.328-333, 1993) in the context of emission tomography, and one is an ad hoc gradient algorithm. The algorithms enjoy desirable local and global convergence properties and combine gracefully with Bayesian smoothing priors. Preliminary numerical testing of the algorithms on simulated data suggest that the convex algorithm and the ad hoc gradient algorithm are computationally superior to the EM algorithm. This superiority stems from the larger number of exponentiations required by the EM algorithm. The convex and gradient algorithms are well adapted to parallel computing
Keywords :
Bayes methods; convergence of numerical methods; emission tomography; image reconstruction; maximum likelihood estimation; medical image processing; smoothing methods; Bayesian smoothing priors; EM algorithm; convex algorithm; convexity argument; emission tomography; exponentiations; global convergence; globally convergent algorithms; gradient algorithm; image reconstruction; local convergence; maximum a posteriori transmission tomography; maximum likelihood algorithms; numerical testing; parallel computing; simulated data; Attenuation; Bayesian methods; Convergence; Image reconstruction; Maximum likelihood estimation; Object detection; Smoothing methods; Stochastic processes; Testing; Tomography;
Journal_Title :
Image Processing, IEEE Transactions on
