一种快速校正CT环形伪影的方法

背景：环形伪影严重影响了CT图像质量，对图像后处理造成困难以及容易造成误诊断。目前去除环形伪影必不可少。目的：去除CT重建图像中的环形伪影，提高CT图像质量以及后续处理和量化分析的精度，便于诊断。方法：首先把含环形伪影的CT图像进行线性变换，将灰度图像转换成浮点类型的图像。接着由直角坐标变换到极坐标，这样原来的环形伪影就被变换成线形伪影。设计多维滤波器，计算每个象素滤波后均值以及方差，通过与闽值比较确定伪影范围。最后通过对伪影范围进行修正以及对图像进行坐标变换，变为灰度图像。结果与结论：通过Matlab7.0软件设计程序，处理含环形伪影的CT图像。实验表明，此方法能有效快速地校正CT环形伪影，是一种属于图像后处理的校正方法。

A new method for quickly correcting ring artifacts in CT image

BACKGROUND： Ring artifacts, seriously affect the CT image quality, create difficulties to post-processing of the images, and easily lead to wrong diagnosis. At present, the removal of ring artifacts is essential. OBJECTIVE： A new image correction method based on polar coordinate transformation is proposed to remove ring artifacts in CT image and to improve the precisions of post processing and quantitative analysis, as well as facilitate the diagnosis. METHODS： Firstly, the CT images with ring artifacts were processed by linear transformation, converting the gray images into floating-point types of images. Secondly, the ring artifacts in the Cartesian coordinate were transformed into the vertical line artifacts in the polar coordinate. Thirdly, designing multidimensional filters, calculated the mean and variance of each pixel after filtering, compared with the threshold value to determine the scope of artifacts. Finally, correcting the range of artifacts, and converting into gray images by coordinate transformation. RESULTS AND CONCLUSION： Designing program by Matlab 7.0, the experiment indicates that this method can correct ring artifacts in CT images quickly and effectively, and it is a image post-processing correction methods.