基于二维经验模式分解的医学图像融合方法  

Medical Images Fusion Algorithm Based on BEMD

作　　者：陈珊[1] 郑琴[1] 秦绪佳[1,2] 

机构地区：[1]浙江工业大学计算机科学与技术学院,杭州310023 [2]浙江省可视媒体智能处理技术研究重点实验室,杭州310023

出　　处：《计算机系统应用》2014年第6期105-110,共6页Computer Systems & Applications

摘　　要：基于多尺度多分辨率的图像融合是医学图像融合的重要方法,二维经验模式分解(BEMD)方法是一种新的多尺度多分辨率图像分解方法.本文提出了一种基于BEMD的医学图像融合方法.首先将待融合的两幅图像进行BEMD分解,获得多个BIMF分量和一个剩余分量;然后针对BIMF分量和剩余分量采用不同的融合规则进行图像融合;最后对融合后的各分量进行BEMD逆变换,得到最终的融合结果.实验结果表明,本文方法可得到较好的融合效果,融合图像清晰,含有的更多信息.Multi-scale and multi-resolution based image fusion is an important fusion method for medical images. The bidimensional empirical mode decomposition （BEMD） method is a new image multi-resolution decomposition approach. A medical image fusion algorithm based on BEMD is presented. Firstly, the two images to be fusined are decomposed into multiple BIMFs components and a residual component by BEMD decomposition. Then, according to the BIMFs and residual components, use different fusion rules for image fusion. Finally, inverse transform each fusioned component of BEMD, the final fusion result is attained. The experimental results show that better fusion result can be attained by using the method, the fusioned image is clearer and more information is contained.

关 键 词：医学图像 图像融合 二维经验模式分解(BEMD) 

分 类 号：TP391.41[自动化与计算机技术—计算机应用技术]