SPECT影像呼吸运动参数辨识与伪影校正的体模研究  

Phantom study for identifying respiratory motion and correcting motion artifacts in SPECT imaging

作　　者：许全盛[1,2] 安晶刚[1,3] 袁克虹[2] 叶大田[1,2] 

机构地区：[1]清华大学医学院生物医学工程系,北京100084 [2]清华大学深圳研究生院生物医学工程研究中心,深圳518055 [3]中国疾病预防控制中心辐射防护与核安全医学所,北京100088

出　　处：《清华大学学报（自然科学版）》2010年第11期1896-1900,共5页Journal of Tsinghua University(Science and Technology)

基　　金：广东省自然科学基金资助项目(No.2008B030303055)

摘　　要：为了提出一种单光子发射断层扫描成像(SPECT)影像呼吸运动参数估计和伪影校正方案,该文通过对伪影图像的方向微分和自相关分析分别辨识出体模的运动方向和幅度,根据匀速运动和正弦振动拟合的系统点扩散函数,运用Wiener滤波和Richardson-Lucy盲反卷积(RL)算法校正伪影。结果表明:5个典型运动方向和幅度辨识的结果均与实际值接近,RL算法校正伪影经过5次迭代后图像质量总体优于维纳滤波。通过图像后处理估计运动参数,运用迭代反卷积可以有效的校正图像伪影,可能成为一种更方便实用SPECT/PET(正电子发射断层扫描成像)图像呼吸伪影解决方案。Respiratory motion identification and corrections for the motion artifacts in SPECT imaging were analyzed in a phantom study by identifying,the blur direction in the SPECT image through a directional derivative analysis of the motion blurred image.The blur extent was estimated by an auto-correlation analysis of the pseudo-whitened blurred image.Then,a Wiener filter and the Richardson-Lucy blind deconvolution algorithm（RL） were employed to restore the motion blurred image according to the estimated point spread function.The results show that the identification results for 5 typical simulated motions are very close to the true value and that the RL algorithm enables better motion artifact correction within 5 iterations than the Wiener filter.This method can be used to identify respiratory motion with a post-reconstruction based deconvolution algorithm as a promising method in clinical practice to correct motion artifacts in SPECT/PET images.

关 键 词：呼吸运动 单光子发射断层扫描成像(SPECT) 运动伪影 方向微分 自相关 反卷积 

分 类 号：TP391.4[自动化与计算机技术—计算机应用技术]