Snake与多尺度分析的医学图像分割研究

摘要

In order to improve the segmentation accuracy, Snake and multi-scale analysis method are combined to segment medical image.This paper makes a rough contour of the image content according to the prior knowledge,processes the image with multi-scale enhancement, and then extracts the contour using snake algorithm in different scales, in this situation the curve is revised in the convergence process, finally extracts the satisfactory contour by optimizing it in different scales.The experimental results dictate the feasibility of the algorithm.In medical image segmentation accuracy,this algorithm is superior to the traditional Snake model.%为提高分割精度,提出Snake与多尺度分析相结合的医学图像分割方法.根据先验知识给定图像一个初始的粗略轮廓,然后对图像进行多尺度增强,在不同的尺度下应用Snake算法进行轮廓提取,相当于在曲线的收敛过程中进行了修正,从而使得轮廓在不同的尺度中逐渐优化,分离出精确的轮廓.实验结果表明,该方法是有效的,对医学图像分割的精度优于传统的Snake模型.