摘要
目前已知的医学图像融合算法未充分考虑源图像间差异性的大小，针对该不足提出了一种基于互信息特征的多模态融合算法。算法引入提升小波变换，将目标图像分解为高、低频子带，根据高频子带的互信息量不同，对低互信息子带采用区域梯度能量与区域标准差相结合的融合规则，对高互信息子带采用边缘强度取大的融合规则。通过多组目标图像融合对比的实验进行验证，算法融合得到的图像信息丰富，边缘清晰，具有良好的视觉特性和优秀的评价指标。
T he know n algorithms of medical image fusion do not consider the difference among different source images . In view of this problem , a multi‐modality fusion algorithm based on mutual information was proposed .The algorithm introduces lifting wavelet transform and decomposes the target image into high frequency sub‐bands and low frequency sub‐bands . According to different mutual information amount of high frequency sub‐bands ,the fusion rule of combining local gradient energy and local standard deviation was adopted for low mutual information sub‐bands ,while fusion rule of bigger edge strength is applied for frequency mutual information sub‐bands .Experimental results of multi‐group target image fusion show the method proposed in this paper is much better because of its rich image information ,clear edge ,good visual features and excellent evaluation indexes .
出处
《浙江理工大学学报（自然科学版）》
2016年第4期607-614,共8页
Journal of Zhejiang Sci-Tech University(Natural Sciences)
基金
国家自然科学基金项目(61374022)
关键词
医学图像融合
提升小波变换
互信息
区域梯度能量
区域标准差
medical image fusion
lifting wavelet transform
mutual information
local gradient energy
local standard deviation
