基于GVF-Snake模型改进的X线图像分割  

Improved X-ray image segmentation based on GVF Snakes model

作　　者：相彬森[1] 

机构地区：[1]东南大学影像科学与技术实验室,南京210096

出　　处：《中国体视学与图像分析》2011年第1期1-4,共4页Chinese Journal of Stereology and Image Analysis

摘　　要：参数活动轮廓模型目前越来越被广泛应用于医学图像分割中,而GVF-Snake模型更是其中研究较为广泛的一种。在人体手部指骨的X线图像分割中,对指节间隙的分割要求较高,传统的GVF-Snake模型方法需要较多的控制点选取,且分割结果并不稳定。因此,本文提出了一种基于改进的GVF-Snake模型分割方法,对图像灰度先进行线性分段处理,再通过直方图的重分布处理,使得图像的对比度增强,便于进行GVF-Snake模型法分割。实验结果表明,采用改进的方法,只需少许几个控制点,即可得到比较精确的图像分割效果,是一种比较好的改进算法。At present,parametric active contour model has been more and more widely used in the segmentation of medical images,and gradient vector flow（GVF） snake model is one of the most widely studied algorithms.In the segmentation of the knuckles,the traditional GVF-Snake model needs some more control points,especially in the gap of the knuckles.In order to get more ideal curves,we propose a new algorithm based on the traditional GVF-Snake model.The improved algorithm firstly processes the image with a piecewise-linear method on the grey level of the image,and then redistributes the histogram to enhance the contrast of the image.The results prove that the new algorithm just needs several control points to get an accurate curve.So we can conclude that the improved algorithm is a better and more suitable algorithm for the segmentation of the knuckles.

关 键 词：图像分割 参数活动轮廓模型 GVF 直方图重分布 

分 类 号：TP391.41[自动化与计算机技术—计算机应用技术]