基于先验图像的CT图像金属伪影校正算法

Metal artifacts correction in CT image based on prior image

目的基于先验图像信息的CT图像金属伪影校正算法可以有效去除CT图像金属伪影。本文对该算法中先验图像模拟生成的投影数据与含有金属伪影的原始CT图像模拟生成的投影数据进行融合过程中的插值方式进行探索研究。方法主要步骤包括:(1)CT图像金属区域分割;(2)前向投影;(3)投影数据金属区域插值;(4)先验图像生成;(5)先验图像模拟生成的投影数据与含有金属伪影的原始CT图像模拟生成的投影数据进行融合;(6)CT图像重建;(7)金属区域填充。在步骤5中,对融合过程中进行的插值运算进行探究,包括多项式拟合的最高次幂和拟合数据量对金属伪影校正后的CT图像质量影响,最后对含有金属伪影的头部CT图像进行本文算法和传统算法在去除金属伪影方面的主观评价和客观指标评价。结果相对于传统算法,本文算法中插值区域两侧各采用3个数据以及进行三次多项式对投影数据中金属区域进行拟合时,CT金属伪影去除效果优质。结论基于先验图像信息的CT图像金属伪影校正算法中的插值计算方法对CT图像金属伪影校正效果产生影响,选择合适的拟合数据量和多项式拟合最高次幂将提高最终校正结果。

Objective The metal artifacts correction method of CT image based on prior image information can effectively reduce metal artifacts,so this paper studies the interpolation during the fusion between prior image projection and original CT image projection.Methods The method includes(1)metal segmentation in CT image;(2)projection operation;(3)interpolation in metal area of sinogram;(4)prior image generation;(5)projection fusion;(6)CT image reconstruction;(7)metal filling.In step 5,the interpolation is investigated in this study,including polynomial maximum order and the amount of sample data used for polynomial fitting.Finally,an head CT image with metal artifacts is used for subjective and objective evaluation.Results When the number of sampler data around the metal area is 3 and the polynomial maximum order is chosen 3,a better result can be obtained.Conclusions The interpolation operation in the metal artifacts correction method of CT image based on prior image information can affect the quality of corrected CT image in terms of metal artifacts.A suitable set of polynomial maximum order and number of sample data used for fitting can improve CT image quality after correction.