摘要： 提出一种基于多分辨率的医学图像区域融合算法。对源图像进行多分辨率分解，考虑邻域系数的影响，基于局部绝对值选大的融合规则，进行高频带的融合。采用模糊C均值算法对熵值大的低频带进行区域分割，将区域映射到其他低频带中，利用EM融合算法融合各个区域。对融合系数进行重构，得到融合图像。实验结果表明，该算法的融合质量较优且效率较高。

关键词: EM融合算法, 多分辨率, 模糊C均值, 区域映射, 邻域系数

Abstract: This paper proposes a medical image region fusion algorithm based on multi-resolution. The source images are decomposed into multi-resolution representations. Considering the influences of the neighborhood coefficients, a local maximum selection rule is proposed to fuse the high frequency bands. The Fuzzy C-means(FCM) algorithm is used to segment the low frequency band with largest entropy into regions which are mapped to the other low frequency bands, then the regions are fused using the EM fusion algorithm. The fused image is obtained by reconstructing the coefficients. Experimental results show that the proposed algorithm has good fusion quality and high efficiency.

Key words: EM fusion algorithm, multi-resolution, Fuzzy C-means(FCM), region mapping, neighborhood coefficient

中图分类号: 

• TP391.41