基于局部能量和PCNN的医学图像融合

An image fusion method which is based on local energy and PCNN

为了满足医学图像对各种组织信息准确性、完整性的要求,本文提出了一种基于局部能量和PCNN的CT和MRI医学图像的融合方法,首先利用非下采样Contourlet变换对待融合的CT图像和MRI图像分别进行多尺度与多方向分解,得到它们各自的低频图像和高频图像,对低频图像,采用局部能量加权的方法进行融合;对高频图像,将它们分别作为脉冲耦合神经网络的输入,再根据它们的神经元的点火次数的大小来进行图像融合。实验结果表明,该算法比以往的融合方法取得了更好融合质量。

An image fusion method which is based on local energy and PCNN is proposed in order to fill the demand of accuracy and integrality.Firstly,the source images are decomposed through nonsubsampled contourlet transform,and then fusions in according of their respective features of low and high frequency.The low frequency exprises smooth regions and uses local energy weighted sums.At the same time,the high frequency includes edges and features information of image that is used as the simplified pulse coupled neural network input links,and then uses PCNN to select high-frequency by the times of fires.The experimental results show that the algorithm is achieving better fusion quality than previous fusion methods.