Abstract
Suppose that each detector unit, d, of an emission scanner measures a count n*(d) which represents the number of emissions into d of an unknown emission density λ.The likelihood, P(n*|λ), is the (Poisson) probability of observing n* under λ.The well-known EM algorithm starts with an estimate λ° of λ and produces a sequence λk, k=l,2,…, of estimators having increasing likelihood and which converge to an estimate λx with maximum likelihood as k→∞. It is well-known that for large k, λk becomes snowy or noisy in appearance, and various methods have been proposed to effectively smooth λk, but these all give up likelihood to get smoothness. We give a method using linear programming of smoothing λk which produces a smoother estimate λk which nevertheless has the same likelihood as λk.
If λk is nearly unique among estimators with its likelihood then λk cannot differ much from λk, while of there are many λ with the likelihood of λk, then λk should be smooth. Experiments described here indicate that for large k, λk is not much smoother than λk, so that λk seems to be nearly unique. This is surprising in view of the fact the problem considered is severely undetermined in the sense that there are typically 3 or more times as many unknowns as equations (measurements), but of course it is the inequality constraints that cause the uniqueness.
We survey the maximum likelihood approach to emission tomography using the iterative EM algorithm and discuss the known difficulty with the algorithm at high iteration numbers.
Access this chapter
Tax calculation will be finalised at checkout
Purchases are for personal use only
Preview
Unable to display preview. Download preview PDF.
References
Baum, L. E. An inequality and associated maximization technique in statistical estimation for probabilistic functions of Markov processes. In Inequalities, III: Proceedings of a Symposium, Oved Shisha ed., Academic Press, New York.
Buchta, C. On nonnegative solutions of random systems of linear inequalities, Discrete Comput. Geom. 2, 1987, pp. 85–95.
Csiszar, I. and Tusnàdy, G. Information geometry and alternating minimization procedures, Math. Inst. of Hungarian Acad. Sci.
Dempster, A. P., Laird, N. M. and Rubin, D. B. Maximum likelihood from incomplete data via the EM algorithm, J. Roy. Stat. Soc. B, 39, 1977, pp. 1–38.
Geman, S. and McClure, D. E. Statistical methods for tomographic image reconstruction, Proc. of 46th Session of ISI, Bull. ISI, vol. 52, 1987.
Johnstone, I. and Silverman, B.W. Speed of estimation in PET, Tech. report 290, March’ 88, Statistics Dept., Stanford University.
Karmarkar, N. K. A new polynomial-time algorithm for linear programming, Combinatorica, 4, 1981, 373–395
Levitan, E. and Herman, G. T. A maximum a posteriori probability expectation maximization algorithm for image reconstruction in emission tomography, IEEE Trans. Med. Imag., 1987.
Liang, Z. and Hart, H. Bayesian image processing of data from constrained source distributions I, II, Bull. Math. Biol. 49, 1987, 51–91.
Llacer, J. and Veklerov, E. The maximum likelihood estimation method of image reconstruction: its fundamental characteristics and their origin, ms., May 1987.
McCarthy, A., Barrett, R. and Miller, M. I. Systolic implementation of the EM algorithm for emission tomograpy on mesh-connected processors, Proc. 22nd Conf. on Inf. Sci. and Systems, Princeton, University, 1988.
Petersen, S. E., Fox, P. T., Posner, M. I., Mintum, M. and Raichle, M. E. Position emission tomographic studies of the cortical anatomy of single-word processing. Nature 331, pp. 585–589, 1988, Feb.
Politte D. G. and Synder, D. L. The use of constraints to eliminate artifacts in maximum-likelihood image estimation for emission tomography, monograph 489, October 1987, Washington University School of Medicine, St. Louis, MO.
Synder, D. L. Random Point Processes, Wiley, New York, 1975.
Shepp, L. A. and Kruskal, J. B. Computerized tomography: The new medical X-ray technology, Amer. Math. Monthly, 85, 420–439.
Synder, D. L. and Miller, M. I. The use of sieves to stabilize images produced with the EM algorithm for emission tomography, IEEE Trans. Nucl. Sci., NS-32, 1985, pp. 3864–3872.
Synder, D. L., Thomas, L. J., Ter-Pogossian, M. M. A mathematical model for positron-emission tomography systems having time-of-flight measurements, IEEE Trans. Nucl. Sci., NS-28, 1981, pp. 3575–3583.
Shepp, L. A. and Vardi, Y. Maximum-likelihood reconstruction for emission tomography, IEEE Trans. Med. Imag., vol. MI-1, pp. 113–122, 1982.
Vanderbei, R. J., Meketon, M., Freedman, B. A modification of Karmarkar’s linear programming algorithm, Algorithmica, 1, 1986, pp. 395–407.
Veklerov, E. and Llacer, J. Stopping rule for the MLE algorithm based on statistical hypothesis testing, ms. February 1987.
Vardi, Y., Shepp, L. A., and Kaufman, L. A statistical model for positron emission tomography, J. Am. Stat. Soc. 80, 1985, pp. 8–37.
Author information
Authors and Affiliations
Editor information
Editors and Affiliations
Rights and permissions
Copyright information
© 1992 Springer-Verlag Berlin Heidelberg
About this paper
Cite this paper
Shepp, L.A., Vanderbei, R.J. (1992). New Insights into Emission Tomography Via Linear Programming. In: Todd-Pokropek, A.E., Viergever, M.A. (eds) Medical Images: Formation, Handling and Evaluation. NATO ASI Series, vol 98. Springer, Berlin, Heidelberg. https://doi.org/10.1007/978-3-642-77888-9_3
Download citation
DOI: https://doi.org/10.1007/978-3-642-77888-9_3
Publisher Name: Springer, Berlin, Heidelberg
Print ISBN: 978-3-642-77890-2
Online ISBN: 978-3-642-77888-9
eBook Packages: Springer Book Archive