基于Mumford-Shah模型的图像快速分割方法  

Fast algorithm of image segmentation based on Mumford-Shah model

作　　者：张博[1] 苏永利[1] 张书玲[1] 

机构地区：[1]西北大学数学系,西安710069

出　　处：《计算机工程与应用》2009年第12期193-194,204,共3页Computer Engineering and Applications

基　　金：陕西省教育厅专项项目No.JK05303~~

摘　　要：针对传统的Mumford-Shah(M-S)模型在图像分割的每一次迭代中都需要另外求解两个椭圆型方程,从而使得算法的执行效率大大降低的缺点,利用分片常函数代替M-S模型中的待优化输出图像,简化了曲线演化的过程,避免了额外计算,有效地提高了算法的执行速度。实验表明,该方法在图像分割效果方面与传统方法没有明显差别,但是程序的运行速度方面却远远优于传统方法。In the process of using traditional Mumford-Shah model in image segmentation,there are always two elliptical equations in each evolution which need to be solved,the solving process seriously influences the speed of algorithm.In order to overcome the shortcoming,a slicing constant function is used to substitute the output image in M-S model,so the calculation is effectively simplified.The experiments show that the algorithm proposed has the same results as the traditional method,but excels in speed.

关 键 词：Mumford—Shah模型 水平集 分片常函数 

分 类 号：TP301[自动化与计算机技术—计算机系统结构]