基于小波变换的CT灌注成像  

CT perfusion imaging based on wavelet transform

作　　者：陈春晓[1] 丁少伟[1] 杨伟[1] 吴苏稼[2] 

机构地区：[1]南京航空航天大学生物医学工程系,南京210016 [2]南京军区南京总医院骨科,南京210002

出　　处：《重庆医学》2012年第3期256-259,共4页Chongqing medicine

摘　　要：目的探讨基于阈值滤波的小波变换用于CT灌注成像的有效性与可行性。方法结合CT灌注成像特性,进行仿真实验,通过峰值信噪比(PSNR)、均方误差(MSE)2个指标来评估不同血流量(BF)及不同方差的高斯噪声下基于小波滤波的灌注模型的效果;对获取的CT骨肿瘤临床数据,运用小波阈值滤波后建立灌注模型,分析肿瘤区域与正常组织的区分情况。结果仿真实验表明,不同BF值及不同方差的高斯噪声下,滤波后PSNR和MSE明显改善;骨肿瘤临床数据表明,本方法具有较强的抗噪声能力,明显区分肿瘤组织与正常组织的边界。结论改进的小波阈值滤波能有效改善CT灌注成像的质量。Objective To investigate the effect and feasibility of CT perfusion imaging using the wavelet transform based on threshold de-noising. Methods In simulated experiments,combining with the characteristics of CT perfusion imaging,the perfusion model based on wavelet filter in different BF and different variance of Gaussian noise is evaluated by parameters of PSNR and MSE; To the clinical data of bone tumors,the model on wavelet threshold de-noising is also used to distinguish the tumor and the normal tissue. Results Simulation experiments show that the PSNR and MSE have obviously been improved in different BF and different variance of Gaussian noise by using the model presented in this paper,and the model also shows powerful abilities of de-noising and keeps edges clearly between normal tissue and tumor. Conclusion The quality of CT perfusion imaging has been effectively im proved by using the model based on the wavelet threshold de-noising,and this study provides more accurate and objective information for doctors.

关 键 词：骨肿瘤 CT灌注成像 小波滤波 仿真实验 

分 类 号：R814.42[医药卫生—影像医学与核医学]