基于改进小波变换的医学图像融合方法  

Medical images fusion method based on revised wavelet transform

作　　者：曲延华[1] 王婷君[2] 薛嗣麟[2] 蒋延友[2] 王安娜[2] 

机构地区：[1]沈阳工程学院自动控制工程系,沈阳110136 [2]东北大学信息科学与工程学院,沈阳110003

出　　处：《计算机应用》2009年第10期2698-2699,2702,共3页journal of Computer Applications

基　　金：东北大学985项目

摘　　要：小波变换在图像融合中的应用已有不少,但大多是热图像和可视图像的融合,在医学图像融合方面的研究还比较少。针对这一现状,提出了基于形态学小波的医学多分辨率融合方法和小波域基于差值图像分割的加权融合方法。该方法保留了小波的视觉效果,同时在很大程度上减少了运算的复杂度,优化了加权系数的计算方法,得到了较好的融合效果。实验结果表明了该算法的优越性。There are many applications of wavelet transform to image fusion. But the images fusioned usually are thermal or visible images. A multi-resolution fusion method based on morphology wavelet and a weighted fusion method based on difference image segmentation in wavelet region for medical images were proposed. These methods maintained the visual effect of wavelet, reduced the computation eomplexity to a great extent, and optimized the computation of weighted coefficients. The experimental results show the excellent fusion effect by these methods.

关 键 词：图像融合 小波变换 数学形态学 多分辨率分析 评价指标 

分 类 号：TP391.41[自动化与计算机技术—计算机应用技术]