Block-iterative Fisher scoring for emission tomography

被引：0
|
作者：
Ma, Jun [1 ]
Hudson, Malcolm [1 ]
机构：
[1] Macquarie Univ, Dept Stat, Sydney, NSW 2109, Australia
来源：
2007 4TH IEEE INTERNATIONAL SYMPOSIUM ON BIOMEDICAL IMAGING : MACRO TO NANO, VOLS 1-3 | 2007年
关键词：
block-iterative Fisher scoring; emission tomography; OS-EM; BSREM; OS-SPS;
D O I：
10.1109/ISBI.2007.356811
中图分类号：
TP18 [人工智能理论];
学科分类号：
081104 ; 0812 ; 0835 ; 1405 ;
摘要：
We introduce and evaluate a block-iterative Fisher scoring (BFS) algorithm for emission tomography. Regularization is achieved by penalized likelihood with a general quadratic penalty. When the algorithm converges, it converges to the unconstrained maximum penalized likelihood (MPL) solution. In a simulated data set, constrained BFS achieves a higher penalized likelihood in fewer iterations than other block-iterative algorithms which are designed for non-negatively constrained penalized reconstruction.
引用
收藏
页码：153 / 156
页数：4
相关论文
共 50 条
12345