基于冗余表达的CT图像中金属伪影校正  

Metal Artifacts Reduction in CT Images via Redundant Representation

作　　者：毛宝林[1] 陈永丽[1] 孝大宇[1] 康雁[1] 

机构地区：[1]东北大学中荷生物医学与信息工程学院,辽宁沈阳110819

出　　处：《东北大学学报（自然科学版）》2013年第1期48-51,共4页Journal of Northeastern University(Natural Science)

基　　金：辽宁省自然科学基金资助项目(201102067);国家自然科学基金资助项目(61071213)

摘　　要：针对CT图像中存在的金属伪影,提出了一种新的基于冗余表达的校正方法.首先分割出CT图像中的金属区域,将其投影得到金属投影区域;然后从原始投影数据中去除金属投影区域部分的投影数据;最后利用冗余表达的算法恢复金属部分的投影数据,经过滤波反投影后,重建出校正后的图像.校正结果显示,本方法对含有多块金属模体的重建图像的金属伪影有较好的校正效果.A new calibration method based on redundant representation was proposed for metal artifacts in CT medical images.First of all,metal projection region was obtained by projecting the segmented metal region.Then,the data of metal projection region was subtracted from the original projection data.Finally,the metal projection data was restored by using redundant representation method and the corrected image was generated after filtered back projection.The calibration results showed that the proposed method can reduce metal artifacts effectively in the reconstructed images.

关 键 词：金属伪影 CT图像 投影数据 滤波反投影 冗余表达 

分 类 号：R445.3[医药卫生—影像医学与核医学]