摘要
目的为快速有效去除CT图像中的环形伪影,本文提出一种基于图像重建后处理的CT环形伪影去除改进方法。方法首先对CT图像从直角坐标系转换成极坐标系,其后对极坐标系下的图像进行滤波,设计一维滤波器,计算每个像素滤波后均值以及方差,通过对计算的方差与方差阈值比较以及该像素点的像素值与像素值阈值比较双重精确确定伪影点位置,并对伪影点进行合理的修正。最后将校正完毕的极坐标图像转换为直接坐标图像。结果仿真和实际CT图像实验表明,本方法在伪影去除、图像保真和计算时间等方面均取得了很好的效果。结论本文提出的方法能够准确识别伪影点位置并有效去除伪影信息,其处理速度能够满足临床要求。
Objective To propose a new method for effectively and rapidly removing the ring artifacts in CT images based on image post-processing, Methods The CT image with ring artifacts in the Cartesian coordinate was first transformed into an image with line artifacts in the polar coordinate. The image in the polar coordinate was then filtered by designing a one-dimensional filter to calculate the mean and variance of each pixel after filtering, which were compared with the variance threshold value and the pixel threshold value to determine the position of the artifacts for corrections accordingly. Finally, the polar coordinate image was converted into Cartesian coordinate image. Results Simulated and actual CT data experimental results demonstrated the efficiency of this method for removing artifacts, retaining the image fidelity and reducing the processing time. Conclusion The new method can accurately recognize the position of the artifacts and effectively remove them to facilitate the clinical diagnosis.
出处
《南方医科大学学报》
CAS
CSCD
北大核心
2012年第12期1748-1751,共4页
Journal of Southern Medical University
基金
国家自然科学基金(30970866)
广州市重大科技专项计划(11BppZLjj2120029)
广东省战略性新兴产业核心技术攻关(2011A081402003)~~
关键词
环形伪影
CT图像
一维滤波器
坐标变换
ring artifacts
CT image
one-dimensional filter
coordinate conversion