一种基于小波变换的动态轮廓模型  

An Active Contour Model Based on Wavelet Transform

作　　者：刘彩霞[1] 

机构地区：[1]枣庄学院计算机科学系,山东枣庄277160

出　　处：《常熟理工学院学报》2009年第2期118-121,共4页Journal of Changshu Institute of Technology

摘　　要：梯度矢量流模型广泛应用于计算机视觉与图像分割,但是受噪声的影响,在处理许多医学图像时不能正确收敛到目标轮廓.本文结合小波变换与梯度矢量流模型提出了一种新的动态轮廓模型.新模型定义小波变换矢量替代梯度矢量,并用多尺度小波变换对图像进行预处理.新模型具有更好的抗噪性,并能够得到连续的边缘点.结果表明新模型提高了动态轮廓模型方法分割复杂图像的能力,是一种有效的医学图像分割方法.GVF snake model is used widely in computer vision and image segmentation, However, there are problems in convergence processing to boundaries of medical image because of noise. A new active contour model based on wavelet transform and GVF （Gradient vector flow） snake model is proposed in this paper. Wavelet transform vector is defined to replace the gradient vector in GVF snake model. And the image is pre-processed with multi-scale wavelet transform. The consecutive edge can be gained and the anti-noise ability is improved with the new model. Experiments indicate that the new algorithm can improve snake model＇s ability to segment complicated image.

关 键 词：动态轮廓模型 梯度矢量流 医学图像处理 图像分割 小波变换 

分 类 号：TP391.4[自动化与计算机技术—计算机应用技术]