图像分割中一种改进的图割模型  

The Modified Graph Cut Model for Image Segment

作　　者：徐利娜[1] 彭国华[1] 

机构地区：[1]西北工业大学应用数学系,西安710072

出　　处：《科学技术与工程》2006年第18期2853-2857,共5页Science Technology and Engineering

摘　　要：RatioCut图割模型仅考虑了像素之间的灰度关系,没有考虑其空间关系,往往会导致分割结果对椒盐噪声及模糊边缘比较敏感,且计算量过大。用流域变换对图像做初始分割,然后建立相应的图割模型,能更有效地利用图像的区域性质,从而增加了分割边界的稳定性,并且由于减少了图中的节点数目,计算时间也大大减少。实验证明了这种模型的优越性。Ratio Cut is a graph cut model. It only considers the gray level relationship among pixels, but not takes into account the spatial relationship among them. It causes that not only the segment result is sensitive to salt-and-pepper noise and blurry edges, but also the quantity of calculation is too large in the algorithm realization. Watershed transform will create the initial image segmentation, then set up the corresponding graph cut model. The method is more effective to make use of the region ptroperty, and the boundary of segment is more stability. This will severely reduce the computational time, due to the reduction of the number of nodes in the graph. The experiment shows superiority of the model.

关 键 词：图像分割 图割模型 RATIO Cut(比例割) 流域变换 

分 类 号：TP391.41[自动化与计算机技术—计算机应用技术]