基于局部占优的自适应像素加权CT/MRI图像融合算法

Adaptive pixel-weighted CT/MRI fusion based on local priority

针对传统基于像素加权平均图像融合算法存在加权因子难以确定的问题,根据CT/MR I图像信息在空间局部互补的特点,提出一种基于局部占优的自适应像素加权算法.根据CT/MR I在空间局部的亮度信息和细节信息的差异,定义并计算局部亮度对比度和局部细节对比度;根据局部亮度和局部细节对比度计算局部占优度矩阵,并进行中值滤波;接着根据局部占优度矩阵计算阈值K,并用阈值K对局部占优度矩阵进行阈值处理,从而得到自适应加权因子矩阵;根据自适应加权因子矩阵对CT/MR I图像进行融合.

There are some difficulties in deciding the weighted factor for the traditional weighted average image fusion. This paper presents an adaptive pixel -weighted method based on CT/MRI local priority. Firstly, we get the local light and local detail contrasts from CT to MRI; Secondly, we get the local priority matrix and make it filtered with middle value filter, based on the local light and local detail contrasts; Thirdly, after having gotten the threshold from the local priority matrix, we process the local priority matrix with threshold and get the adaptive weighted factor matrix; Finally we fuse the CT/MRI image with the adaptive factor matrix.