Catalogo Articoli (Spogli Riviste)
OPAC HELP
Titolo: FAST AND STABLE MAXIMUM APOSTERIORI CONJUGATEGRADIENT RECONSTRUCTION ALGORITHM
Autore: LALUSH DS; TSUI BMW;
 Indirizzi:
 UNIV N CAROLINA,DEPT BIOMED ENGN,CAMPUS BOX 757,152 MACNIDER HALL CHAPEL HILL NC 27599 UNIV N CAROLINA,DEPT RADIOL CHAPEL HILL NC 27599
 Titolo Testata:
 Medical physics
fascicolo: 8,
volume: 22,
anno: 1995,
pagine: 1273  1284
 SICI:
 00942405(1995)22:8<1273:FASMAC>2.0.ZU;29
 Fonte:
 ISI
 Lingua:
 ENG
 Soggetto:
 POSITRON EMISSION TOMOGRAPHY; GIBBS PRIOR DISTRIBUTIONS; MODIFIED EM ALGORITHM; IMAGERECONSTRUCTION; DETECTOR RESPONSE; ITERATIVE RECONSTRUCTION; ATTENUATION CORRECTION; LIKELIHOODESTIMATION; SPECT RECONSTRUCTION; COMPENSATION;
 Keywords:
 ITERATIVE RECONSTRUCTION; SPECT RECONSTRUCTION;
 Tipo documento:
 Article
 Natura:
 Periodico
 Settore Disciplinare:
 Science Citation Index Expanded
 Citazioni:
 39
 Recensione:
 Indirizzi per estratti:



 Citazione:
 D.S. Lalush e B.M.W. Tsui, "FAST AND STABLE MAXIMUM APOSTERIORI CONJUGATEGRADIENT RECONSTRUCTION ALGORITHM", Medical physics, 22(8), 1995, pp. 12731284
Abstract
We have derived a maximum a posteriori (MAP) approach for iterative reconstruction based on a weighted leastsquares conjugate gradient (WLSCG) algorithm. The WLSCG algorithm has been shown to have initial convergence rates up to 10x faster than the maximumlikelihood expectation maximization (MLEM) algorithm, but WLSCG suffers from rapidly increasing image noise at higher iteration numbers. In our MAPCG algorithm, the increasing noise is controlled by a Gibbs smoothing prior, resulting in stable, convergent solutions. Our formulation assumes a Gaussian noise model for the likelihood function. When a linear transformation of the pixel space is performed (the ''relaxation'' accelerationmethod), the MAPCG algorithm obtains a lownoise, stable solution (one that does not change with further iterations) in 1030 iterations, compared to 100200 iterations for MAPEM. Each iteration of MAPCG requires approximately the same amount of processing time as one iteration of MLEM or MAPEM. We show that the use of an initial image estimate obtained from a single iteration of the Chang method helps the algorithm to converge faster when acceleration is not used, but does not help when acceleration is applied. While both the WLSCG and MAPCG methods suffer from the potential for obtaining negative pixel values in the iterated image estimates, the use of the Gibbs prior substantiallyreduces the number of pixels with negative values and restricts them to regions of little or no activity. We use SPECT data from simulated hatsphere phantoms and from patient studies to demonstrate the advantages of the MAPCG algorithm. We conclude that the MAPCG algorithm requires 10%25% of the processing time of EM techniques, and provides images of comparable or superior quality.
ASDD Area Sistemi Dipartimentali e Documentali, Università di Bologna, Catalogo delle riviste ed altri periodici
Documento generato il 20/01/20 alle ore 16:01:28