基于NSST和改进PCNN的医学图像融合

Medical image fusion based on NSST and improved PCNN

为了解决单一模态医学图像的局限性,提出了一种基于非下采样剪切波变换(NSST)和改进型脉冲耦合神经网络(PCNN)相结合的多模态医学图像融合方法。首先,利用NSST对源图像进行多尺度、多方向分解,得到低频子带系数和高频子带系数;其次,低频子带系数由区域能量和方差求取区域特征,采用基于区域特征加权的方式进行融合;高频内层子带系数先通过PCNN求出区域点火特性,再与平均梯度加权的方式进行选择,高频外层子带系数采用区域绝对值取大的融合规则;最后,通过逆NSST重构图像。实验结果表明:与常用融合规则对比,在主观效果上,本文的融合图像可以保留源图像的边缘信息,得到更好的视觉效果;在客观指标上,本文方法融合得到的图像在互信息(MI)、边缘评价因子(QAB/F)和结构相似度(SSIM)等客观评价指标上取得更好的效果。

In order to solve the limitation of single mode of medical images,a multimodal medical images fusion algorithm based on non-subsampled shearlet transtransform （NSST） and improved pulse coupled neural network （PCNN） is proposed. First, the source medical images are decomposed into low and high frequency subbands by NSST. Moreover, regional characteristics of the low frequency subband coeffi- cients are obtained from regional energy and variance, a fusion rule based on weighted regional character- istics is adopted for the low frequency subband coefficients; and with the firing times and average gradient weighted method,the inner coefficients of the high frequency bands of medical images after decomposition are fused, and the firing times are determined by the improved PCNN;the outer coefficients of the high frequency bands are combined by the maximal regional absolute value. Finally,the inverse NSST is used to produce the fused image. It is found through image fusion experiment that, subjectively, the image fused by the proposed algorithm preserves marginal information of source images effectively, and has better objective visual quality,objectively,fusion image of the new fusion rule is more superior in the objective evaluation index, such as mutual information （MI）, edge evaluation factor （QAB/F） and structural similarity （ SSIM）.