東華大學圖書館 |

語系: 繁體中文

說明(常見問題)

回圖書館首頁

手機版館藏查詢

登入

回首頁

切換: 標籤 | MARC模式 | ISBD

Fast algorithms for PET image recons...

Li, Quanzheng.

FindBook

Google Book

Amazon

博客來

Fast algorithms for PET image reconstruction and their analysis .

紀錄類型:	書目-電子資源 : Monograph/item
正題名/作者:	Fast algorithms for PET image reconstruction and their analysis ./
作者:	Li, Quanzheng.
面頁冊數:	118 p.
附註:	Source: Dissertation Abstracts International, Volume: 67-10, Section: B, page: 5951.
Contained By:	Dissertation Abstracts International67-10B.
標題:	Engineering, Biomedical. -
電子資源:	http://pqdd.sinica.edu.tw/twdaoapp/servlet/advanced?query=3237165
ISBN:	9780542912719

Fast algorithms for PET image reconstruction and their analysis .
Li, Quanzheng.

Fast algorithms for PET image reconstruction and their analysis . - 118 p.

Source: Dissertation Abstracts International, Volume: 67-10, Section: B, page: 5951.

Thesis (Ph.D.)--University of Southern California, 2006.

Iterative Bayesian image reconstruction can achieve superior image quality compared to analytic image reconstruction by applying an accurate system model and an appropriate data model. However, due to the increased number of detectors in modern PET scanners and higher image resolution required by new applications, the amount of computation in image reconstruction increases faster than computer power, and consequently, conventional iterative Bayesian image reconstruction algorithms are often impractical. Furthermore, the decrease of scan duration in clinical and dynamic reconstruction leads to low count rate data, which deems the traditional data models and analysis methods inappropriate. This dissertation focuses on these challenges to develop more accurate and faster methods for static and dynamic PET image reconstruction.

ISBN: 9780542912719Subjects--Topical Terms:

1017684
Engineering, Biomedical.

Fast algorithms for PET image reconstruction and their analysis .
LDR:03247nmm 2200313 4500 001 1833110
005 20070907105142.5
008 130610s2006 eng d
020 $a 9780542912719
035 $a (UMI)AAI3237165
035 $a AAI3237165
040 $a UMI $c UMI
100 1 $a Li, Quanzheng. $3 1921822
245 1 0 $a Fast algorithms for PET image reconstruction and their analysis .
300 $a 118 p.
500 $a Source: Dissertation Abstracts International, Volume: 67-10, Section: B, page: 5951.
500 $a Adviser: Richard M. Leahy.
502 $a Thesis (Ph.D.)--University of Southern California, 2006.
520 $a Iterative Bayesian image reconstruction can achieve superior image quality compared to analytic image reconstruction by applying an accurate system model and an appropriate data model. However, due to the increased number of detectors in modern PET scanners and higher image resolution required by new applications, the amount of computation in image reconstruction increases faster than computer power, and consequently, conventional iterative Bayesian image reconstruction algorithms are often impractical. Furthermore, the decrease of scan duration in clinical and dynamic reconstruction leads to low count rate data, which deems the traditional data models and analysis methods inappropriate. This dissertation focuses on these challenges to develop more accurate and faster methods for static and dynamic PET image reconstruction.
520 $a We develop a simple but accurate statistical data model by exploiting the symmetry of the probability mass function of the data. We derive an approximate log-likelihood function using our new statistical data model and develop a new reconstruction method. Simulation results show the approximation improves on earlier methods and the reconstruction results have good bias performance even at very low count rates. Further, we present a method to accurately estimate the Fisher information matrix (FIM), which plays a key role in the analysis of reconstructed images and cannot be precisely computed by conventional methods when the count rate is small. We also demonstrate the resulting FIM can be applied to control the reconstructed image quality.
520 $a To reduce the run time caused by increasing computation load, we apply the incremental optimization transfer method to static PET image reconstruction and develop a fast hybrid reconstruction algorithm using it with incremental gradient methods and conjugate gradient methods. Through automatic selection of the point at which we switch between these three algorithms, we achieve faster convergence than individual algorithms. For dynamic reconstruction, we describe a fast globally convergent fully 4D incremental gradient algorithm to estimate the continuous-time tracer density from list mode PET data. Fully 4D simulations demonstrate the convergence of the algorithm for a high count data set on a 4-ring scanner.
590 $a School code: 0208.
650 4 $a Engineering, Biomedical. $3 1017684
650 4 $a Engineering, Electronics and Electrical. $3 626636
650 4 $a Health Sciences, Radiology. $3 1019076
690 $a 0541
690 $a 0544
690 $a 0574
710 2 0 $a University of Southern California. $3 700129
773 0 $t Dissertation Abstracts International $g 67-10B.
790 1 0 $a Leahy, Richard M., $e advisor
790 $a 0208
791 $a Ph.D.
792 $a 2006
856 4 0 $u http://pqdd.sinica.edu.tw/twdaoapp/servlet/advanced?query=3237165