改进的Snake算法在颅脑CT图像分割中的应用

Application of Improved Snake Algorithm in Segmentation of CT Brain Images

尝试用一种改进的Snake算法创建模型,自动分割CT头颅断层图像的脑组织和脑脊液,用于后期辅助医生对患病部位的诊断和治疗,为临床提供更方便的工具。考虑到脑部CT图像和相邻的骨组织灰度差异较大,利用改进Snake算法对CT头颅图像进行自动分割,将分割的脑组织图像与专业医学影像医生手工分割的CT脑组织图像做对比,发现这种改进的Snake算法分割的CT颅脑图像从轮廓外形、灰度均值和方差值都和专业医生手工分割的CT颅脑图像非常接近,表明这种改进的Snake算法分割CT颅脑图像是可行的。

Through an improved Snake algorithm a model was established for the automatic segmentation of the CT image of brain tissues and cerebrospinal fluid to assist the clinical diagnosis and treatment of cerebral diseases. In view of the differences between the CT image of brain and grayscale of adjacent bone tissues,the CT brain image was automatically segmented through the improved Snake algorithm,and the images automatically segmented and those manually segmented by the doctor. The results showed that the contour,grayscale mean and variance of the two segmented images were very close. Thus,the segmentation of CT brain images through the improved Snake algorithm was feasible.