脊椎图像分割与椎体信息提取的算法及应用  

Application of spine images segment and extract information

作　　者：李滚[1] 刘欢[1] 柯善群 

机构地区：[1]西安工业大学电子信息工程学院,陕西西安710021

出　　处：《电子设计工程》2017年第16期183-186,共4页Electronic Design Engineering

基　　金：西安工业大学校长基金(XAGDXJJ14011)

摘　　要：针对传统阈值分割算法的一些缺点,通过将数字形态学与阈值分割算法相互结合提出了一种改进的阈值分割算法来进行脊椎图像分割,并将分割结果与传统图像分割方法得到的结果进行分析对比。结果显示本论文改进的阈值分割算法得到的脊椎分割图像与传统方法分割的图像相比在抑制噪声方面有好的效果,在对比度方面也优于其他传统的医学图像分割方法,最后设计了计算脊椎相邻椎体间的相对距离、相对椎体厚度及其变化率的有效算法,可为临床进一步应用提供参考。In view of some disadvantages of the traditional threshold segmentation algorithm, an improved threshold segmentation algorithm is proposed by combining the digital morphology and the threshold segmentation algorithm, the results show that the improved threshold segmentation algorithm to segment after spine image compared to the segmentation of the traditional methods of image. In a large extent reduces over segmentation phenomenon and in the aspect of noise suppression also outperforms other traditional medical image segmentation method. Finally, the distance between the vertebrae, the spine adjacent vertebrae thickness and its change rate are all calculated, all of them provide for further clinical reference.

关 键 词：阈值分割 医学图像 图像分割 光学图像处理 

分 类 号：TN911.73[电子电信—通信与信息系统]