基于多级混合模型的图像分割方法

本文研究了典型的基于区域的图像分割方法主动形状模型（Active Shape Model, ASM）和基于边缘的图像分割snake 算法，分析了算法适用条件和各自的优缺点。结合snake 模型与主动形状模型，提出了一种基于多级混合模型的图像分割方法：首先使用ASM，获得的对目标图像位置和形态的最佳估计，并通过snake 算法对匹配的结果进行二次调整，获取可以精确描述图像边缘信息的模型。试验表明了该方法的有效性。
关键字：主动形状模型；蛇模型；图像分割

Abstract: Based on the analysis of Active Shape Model(ASM)and snake algorithm ,which gives us the application conditions and the advantages and disadvantages, A multistage hybrid matching method by integrating statistical model and snake is presented for image segmentation. A statistical model can locate objects rapidly, then a snake by adding an additional terms for the target object is built in order to consider the smoothness of the target contour.The method is well suited for automatic image segmentation.
Keywords: Active Shape Model; snake algorithm; image segmentation