基于改进平滑硬阈值约束的CT迭代重建算法  

Iterative CT Image Reconstruction Based on Modified Smooth Hard Threshold Regularization

作　　者：骆毅斌 陈梓嘉[2] 徐月晋 胡洁 

机构地区：[1]广州华端科技有限公司,广州510530 [2]南方医科大学生物医学工程学院医疗仪器研究所,广州510515

出　　处：《核电子学与探测技术》2015年第9期940-943,共4页Nuclear Electronics & Detection Technology

基　　金：国家科技支撑计划课题(2015BAI01B10)资助

摘　　要：稀疏角度投影数据重建CT图像,是一种有效的降低X射线辐射剂量的方式。提出一种改进的平滑硬阈值(SHT)约束,使其更接近于L_0范数,同时最优参数不受重建图像影响。该重建算法沿用ART-TV的重建策略,即对每次ART重建后的图像进行改进SHT的最小化约束。实验表明:新算法提高了重建图像质量,信噪比更高,平均绝对误差更小。Image reconstruction for sparse view computed tomography is an effective way to reduce x - ray radia- tion. A modified SHT has been proposed for sparse - view CT image reconstruction. The improved SHT is much closer to L0 norm than SHT. The parameter in modified SHT is not affected by reconstructed image. Similar to ART- TV reconstruction strategy, our method in this study includes ART reconstruction followed by improved SHT minimization. Experimental results have demonstrated that new method can lead to images with better qual- ity, higher signal - to - noise ratio and lower mean absolute error.

关 键 词：稀疏角度 CT图像 迭代重建 平滑硬阈值 L0范数 

分 类 号：TP391.1[自动化与计算机技术—计算机应用技术]