新型非局部平均滤波修正的CT图像重建算法  

CT image reconstruction algorithm based on novel nonlocal means filter

作　　者：胡洁 陈梓嘉[2] 骆毅斌 

机构地区：[1]广州华端科技有限公司,广东广州510670 [2]南方医科大学生物医学工程学院,广东广州510515

出　　处：《计算机工程与设计》2016年第10期2780-2783,2788,共5页Computer Engineering and Design

基　　金：国家科技支撑计划课题基金项目(2015BAI01B10)

摘　　要：为解决基于传统非局部平均滤波修正的CT图像重建算法中,重建图像结构边缘过于平滑的现象,提出将新型非局部平均滤波用于CT图像重建。对传统的非局部平均滤波算法中的相似窗定义和滤波参数进行改进,得到新型非局部平均滤波算法。重建算法包括代数重建算法、非负约束、新型非局部平均滤波,以上3步循环操作,直到满足收敛条件。对Shepp-Logan体模进行仿真重建实验,实验结果表明,与传统方法相比,改进算法重建出的图像信噪比更高,边缘保持更好。To solve the problem that CT image reconstruction method based on conventional non-local means（NLM）algorithm leads to over-smoothed image,a novel NLM algorithm was proposed.The novel NLM algorithm was developed by defining new similarity window and filtering parameter.The proposed CT image reconstruction included algebraic reconstruction technique（ART）,positivity constraint,novel NLM filtering.The above processes were repeated until the convergence criterion was satisfied.The restults of simulation on Shepp-Logan show that the proposed method can reconstruct image with better signal to ratio（SNR）and structural preserving effects,compared to conventional CT image reconstruction algorithm.

关 键 词：计算机断层摄影 非局部平均 相似窗 滤波参数 迭代重建 

分 类 号：TP391.41[自动化与计算机技术—计算机应用技术]