基于NSST与引导滤波的医学图像融合算法  

Medical image fusion algorithm based on NSST and guided filtering

作　　者：石平霞 陈世国[1] 丁冬冬 SHI Pingxia;CHEN Shiguo;DING Dongdong(College of Physics and Electronics Science,Guizhou Normal University,Guiyang 550025,China)

机构地区：[1]贵州师范大学物理与电子科学学院,贵州贵阳550025

出　　处：《传感器与微系统》2022年第5期131-134,共4页Transducer and Microsystem Technologies

基　　金：贵州省科学技术基金资助项目(黔科合J字[2010]2145号)。

摘　　要：针对目前医学图像融合领域中,融合图像存在细节信息不够清晰及信息丢失的问题,提出一种基于非下采样剪切波变换(NSST)与引导滤波相结合的医学图像融合算法。首先,对已经过严格配准的两幅医学源图像进行NSST,得到各自的低频子带与高频子带。其次,低频子带采用基于引导滤波的加权平均的融合方法,高频子带采用平均梯度、区域能量指导加权系数的融合规则。最后,将低频子带和高频子带进行非下采样剪切波反变换,从而得到最后的融合图像。最终的实验结果表明:与文中所提到的几种融合算法相比,该算法不仅在多项评价指标中表现更佳,而且也有较好的视觉效果。Aiming at the problem that in current medical image fusion field,detailed information of the fused image is not clear enough and the information is lost,a medical image fusion algorithm based on the combination of non-subsampled shear wave transformation(NSST)and guided filtering is proposed.Firstly,the two medical source images that have been strictly registered are subjected to NSST to obtain respective low-frequency subbands and high-frequency subbands.Secondly,the low-frequency subband uses a weighted average fusion method based on guided filtering,and the high-frequency subband uses a fusion rule of average gradient and regional energy to guide the weighting coefficient.Finally,the low-frequency sub-band and the high-frequency sub-band are inversely transformed by NSST to obtain a fused image.The final experimental results show that,compared with several fusion algorithms mentioned in this article,the algorithm not only performs better in multiple evaluation indexes,but also has better visual effects.

关 键 词：图像融合 医学图像 非下采样剪切波变换 引导滤波 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]