一种多小波滤波器在医学图像融合中的应用  

Application of multiwavelet filter in medical image fusion

作　　者：王修信[1,2,3,4] 胡维平[1] 李永军[5] 

机构地区：[1]广西师范大学计算机科学与信息工程学院 [2]遥感科学国家重点实验室,北京100875 [3]环境遥感与数字城市北京市重点实验室 [4]北京师范大学地理学与遥感科学学院遥感与地理信息系统研究中心 [5]河南大学物理与信息光电子学院,河南开封475001

出　　处：《计算机工程与应用》2007年第15期198-200,214,共4页Computer Engineering and Applications

基　　金：国家自然科学基金(the National Natural Science Foundation of China under Grant No.40571109) ;广西科学基金( 桂科自 0339036) 。

摘　　要：医学图像融合为临床诊断、手术设计与效果评估等提供不同模态影像互补信息。构造具有3阶消失矩的双正交多小波滤波器,其以高斯函数一次泰勒规范展开正交化后得到的完备二次单位正交基作为尺度函数。利用构造的多小波滤波器,分别对CT、MRI医学图像进行分解,使用基于多小波分解改进的局部能量加权规则融合图像。结果表明改进规则融合了源图像主要信息,优于常用的融合规则。应用构造的多小波滤波器可有效地融合医学图像,略优于DGHM多小波。Medical image fusion provides complementary information of the different modal images for diagnosis,surgery design and evaluation.Therefore,complete unit orthogonal bases with second degree were obtained by orthonormalizing first order Taylor expansion of Gauss function.The biorthogonal muhiwavelet filter with third order vanished moment was designed with the bases as scale functions.First CT and MRI medical images were discomposed with the designed filter.Then the fused scheme with local energy weight was modified on the multilevel.Finally the fused image was obtained with the scheme.The result indicated the fused image preserved the details of each input image successfully.h was more effective than the general fused scheme.The designed filter can be used to fuse medical image efficiently,and is a little effective than DGHM multiwavelet.

关 键 词：医学图像 滤波器构造 多小波变换 图像融合 多层能量加权 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]