摘要
为解决现有医学图像融合算法存在的图像失真和实时性差等问题,提出一种在非下采样剪切波变换(NSST)域内将压缩感知(CS)和脉冲耦合神经网络(PCNN)相结合的融合算法。该算法利用NSST将MRI图像和CT图像分解成高频系数和低频系数。采用CS与PCNN相结合的算法融合高频系数,将源图像压缩采样得到观测值,作为PCNN神经元的反馈输入;使用区域特征加权融合低频系数;最后通过逆NSST变换获得融合图像。实验证明,该算法有效保留了源图像的细节信息,融合图像的分辨率和亮度也有所提升,并在空间频率、标准差和边缘评价因子等客观评价指标上取得良好的效果,减少了运行时间。
A fusion algorithm that combines compressed sensing(CS)and pulse-coupled neural network(PCNN)in the domain of non-subsampled shearlet transform(NSST)is proposed to solve the problem of image distortion and poor real-time performance in the existing medical image fusion algorithm.NSST is used to decompose MRI image and CT image into high frequency coefficient and low frequency coefficient,among which the high frequency coefficient is fused by the algorithm that combines CS and PCNN.The source image is compressed and sampled to obtain the observation value,which is taken as the feedback input of PCNN neuron.The low frequency coefficient is fused by regional feature weighting.Finally,the fusion image is obtained by inverse NSST transformation.The experimental results show that the algorithm can effectively preserve the details of the source image and improve the resolution and brightness of the fusion image.Moreover,running time has been reduced and good results have been achieved in objective evaluation indexes such as spatial frequency,standard deviation and edge evaluation factor.
作者
罗欣
吴亚娟
余晓江
乐晓飞
LUO Xin;WU Yajuan;YU Xiaojiang;YUE Xiaofei(College of Computer Science,China West Normal University,Nanchong Sichuan 637009,China)
出处
《西华师范大学学报（自然科学版）》
2020年第1期111-116,共6页
Journal of China West Normal University(Natural Sciences)
基金
西华师范大学英才科研基金项目(17YC163)。
关键词
医学图像融合
非下采样剪切波变换
脉冲耦合神经网络
压缩感知
区域特征加权
medical image fusion
non-subsampled shearlet transform
pulse-coupled neural network
compressed sensing
region feature weighting
