中文摘要：

该文提出了一种基于图像配准和分割的方法来对脑白质疏松症MR图像中脑白质病变区域进行量化。首先,利用标记分水岭法在T1flair序列上将头骨等组织去除,并利用多尺度小波变换法提取出脑白质区域;其次,通过对T1和T2flair序列进行配准,提取出T2flair序列上脑白质区域,并采用Otsu阈值法,在相应的T2序列脑白质区域图上分割出脑白质病变区域;最后,通过分割后计算得到的脑白质体积值和病变区域的体积值,计算出病变区域所占百分比。结果表明,该方法实现了对脑白质疏松症MR图像的计算机自动分割,能够衡量脑白质疏松症的病变程度。因此,此方法在脑白质疏松症辅助诊断上具有临床应用价值。

英文摘要：

To analyze quantitatively of the lesion areas in leukoaraiosis MR image,a method of registration and segmentation was presented in this paper.First,the skull portion of MR image on T1 flair sequence was removed by marker-driven watershed algorithm,then multi-scale wavelet transform method was applied to extract the white matter regions;Second,the white matter regions on T2 flair sequence would be extracted by image registration of T1 T2 flair,and then by using Otsu threshold method,the white matter lesions would be separated from the corresponding T2 white matter area;Finally,the volume percentage of white matter lesions would be calculated from the segmented white matter volume and white matter lesion volume.Results show this method can realize leukoaraiosis＇ MR image segmentation automatically,which can further evaluate the severity of leukoaraiosis.Therefore,this method has clinical value on assistant diagnosis to leukoaraiosis.