D. S. Lalush and B. M. W. Tsui, "Attenuation and detector response compensations used with gibbs prior distributions for maximum a posteriori SPECT reconstruction," IEEE Nuclear Sci. Symposium and Med. Imaging Conf., vol. 3, pp. 1797--1802, 1991. Home/Search Document Not in Database Summary Related Articles Check |
This paper is cited in the following contexts:
Effects of Attenuation and Blurring in Cardiac SPECT and.. - Di Bella (1995) (Correct)
....the signal (the variance is equal to the mean of the signal) Additive noise is an approximation if the process is Poisson. That is, find the object values x that are most likely to generate the observed data y. Research is currently being done on the application of MAP to emission tomography [22, 23, 24]. MAP may be viewed as applying a constraint to the inverse problem. As discussed further in Section 2.6.5, it seems agreed upon that a great deal of care must be given to choosing the priors (p(x) in order not to bias the solution. Also, as pointed out by Hanson [24] there is a philosophical ....
....in PET. Chapters 5 and 6 propose ways to speed the 3 D reconstruction by using small kernels recursively and by using parallel processing. 2.6. 7 Comparison with Filtered Backprojection For SPECT with simulated data, ML EM without any compensations is thought by many in the research community [22, 38, 39, 40, 41] to be comparable or marginally better than filtered backprojection, in terms of resolution, noise, and image quality. In low count images, ML EM is thought to have more of an advantage over filtered backprojection, with the explanation being that while filtered backprojection spreads the noise ....
D. S. Lalush and B. M. W. Tsui, "Attenuation and detector response compensations used with gibbs prior distributions for maximum a posteriori SPECT reconstruction," IEEE Nuclear Sci. Symposium and Med. Imaging Conf., vol. 3, pp. 1797--1802, 1991.
Using Local Median as the Location of the Prior.. - Alenius.. (2 citations) (Correct)
....image [3] The prior is supposed to describe the nature of the true radioactivity concentration of the tissue. Deviations of the image being reconstructed from this prior are penalized. These methods may require non trivial and case by case tuning of multiple parameter values of the prior [4] [5]. Priors that minimize an energy term of a neighborhood of pixels act as roughness penalties [6] They tend to blur edges because both noise and edges contribute to high energy 1 http: www.cs.tut.fi sakkeus mrp mic97m1143.ps. This is the MIC97 Albuquerque conference record preprint. Copyright ....
....resulting image provides a good fit with the measured data. B. Bayesian Methods In order to suppress the noise of the MLEM method a constraint can be used. The Bayesian MAP approach maximizes the a posteriori probability: f(jn) f(nj)f( A general Gibbs form for the a priori probability is [5] f( Ae GammafiU ( 4) where fi is the weight of the prior. The non negative function U( has its minimum when the image meets the prior assumptions. It can be an energy function computed according to differences between neighboring pixels [6] The complete data formulation and the E ....
[Article contains additional citation context not shown here]
D. Lalush, B. Tsui "Attenuation and detector response compensations used with Gibbs priors distributions for maximum a posteriori SPECT reconstruction,"IEEE Transactions on Nuclear Science, vol. 39, pp. 1454-1459, Oct 1992.
Online articles have much greater impact More about CiteSeer.IST Add search form to your site Submit documents Feedback
CiteSeer.IST - Copyright Penn State and NEC