【摘要】 在计算机断层成像(computer tomography,CT)技术中,散射和射束硬化等因素会使CT图像出现放射状的金属伪影.这些金属伪影影响着医生对疾病的诊断同时干扰放射治疗过程中的剂量计算等.为了解决目前临床实践中广泛使用的CT扫描机成像中的金属伪影问题,本文就牙科金属材料形成的伪影问题提出了一种基于极坐标变换的校正方法.该算法通过极坐标变换将放射状伪影条纹转换为行列可分的线形伪影,并对得到的线形伪影图像进行频域滤波处理,最终得到校正后的CT图像.采用标准差和平均CT值差异等参数对11例患者CT图像数据进行量化研究,结果表明处理后的CT图像伪影条纹减少.该算法能够在一定程度上去除CT图像中放射状的伪影条纹,并且保护了图像边缘信息.更多还原

【Abstract】 In computer tomography( CT) techniques,scattering and beam hardening may render radial artifacts in CT images that affect disease diagnosis,interfering with the dose calculation in radiotherapy and so on. To minimize the impact of metal artifacts,a new method based on polar coordinate transformation is proposed for the correction of X-ray metal artifacts that arise from dental fillings. The proposed algorithm converts radial artifacts into linear artifacts through polar coordinate transformation. Then a two dimensional image filtering in the frequency domain is designed to filter out the linear artifacts. Quantitative study with data of 11 patients applying the standard deviation and the average CT value difference shows that the proposed algorithm can reduce the radial metal artifacts in the CT image to a certain extent,and retain the edge information of images as well.更多还原