文档介绍：孙守思,等:离散重建点模型的EM迭代重构 433 0 离散重建点模型的EM迭代重构孙守思1’2?,邱钧3,桂志国1’2, 刘畅3 (, 太原030051; , 太原030051 , 北京 100101) 摘要:发射式断层成像中,EM算法建立在传统离散化模型下,用观测数据来估计未知图像。基于像素格的传统的离散化模型,简单直观,但对Radon变换的离散化存在误差。本文引入离散重建点模型。减小了离散化模型带来的误差,使重建过程更加准确。建立在新模型下的EM算法,有更快的收敛速度。关键词:图像重建;重建点;离散化模型;EM算法文献标识码:A The EM iterative reconstruction of ediscrete reconstruction points SUN Shousil·2”,QIU Jun3,GUI Zhigu01”,LIU Chan93 ( Key Laboratory forElectronic Measurement Technology,North University ofChina, Taiyuan 03005 1,China; Laboratory ofInstrumentation Science&Dynamic Measurement, North University ofChina,Taiyuan 03005 1,China; ofApplied Mathematics,Beijing Information Science&Technology University,Beijing 100101,China) Abstract:In emission tomography,the EM algorithm establishing under traditional discrete model uses observed data toestimate theunknown traditionaldiscretemodel based on pixels issimple and intuitive,but there existsdeviations indiscretization toRadon paper introduces amodel of discrete reconstruction reduces deviations caused by the discretemodel and makes the recon— struction more EM algorithm inthis new model has fasterconvergent speed. Key words:Tomography;Imaging Point;Discrete Model;EM Algorithm; 引言发射式断层成像中,核素在被测物内发生衰变时释放出正电子,正电子与周围电子发生湮灭时发射出光子对。通过泊松概率模型,来描述放射性物质发射粒