基于GVF模型的图像分割方法的改进  

Improvement of image segmentation based on GVF model

作　　者：王海军[1] 张有志[1] 

机构地区：[1]山东大学信息科学与工程学院,山东济南250100

出　　处：《计算机应用》2006年第5期1040-1041,1044,共3页journal of Computer Applications

摘　　要：在对梯度矢量流(GradientVectorFlow,GVF)模型的力场分析的基础上,将力场扩散方程中梯度矢量流的各分量分别归一化,使曲线上各点受物体边界力场影响跟它与边界的距离无关,从而克服了GVF模型难以解决的深凹腔问题。同时,提出在GVF模型中引入和设置方向矢量的算法,在矢量指定范围内能屏蔽掉不需要的物体力场的影响,从而更加准确地分割出感兴趣的物体。Based on the analysis of the force field of gradient vector flow （GVF） model, the force vectors of GVF in the diffusion equations were normalized, which made the external forces irrelevant of the distance between the points of the contour and the boundary of object. Thus this method can solve deep boundary concavities. At the same time, by introducing direction vector into the GVF model, in the region specified by this vector, the force fields of the objects not wanted can be shielded, and the interested objects can be segmented more precisely.

关 键 词：图像分割 SNAKE模型 GVF模型 方向矢量 

分 类 号：TP391.41[自动化与计算机技术—计算机应用技术]