基于互信息的医学图像匹配中的改进插值算法

Expanded Interpolation in Medical Image Registration Based on Mutual Information

基于互信息的配准方法是医学图像配准领域的重要方法,具有鲁棒性,精度高等优点,已成为医学图像处理领域的热点。在计算两个图像之间的互信息时,为了图像配准精度,图像的像素点经过空间变换需要进行插值,目前采用的插值方法一般都定位在单一的传统插值算法,由于传统的插值算法存在插值精度低或插值速度慢的缺点,提出了一种基于像素点的亮度绝对误差的图像插值算法,插值算法结合了近邻插值算法和双三次插值算法的优点,提高了配准的速度和精确度。通过对头部图像进行配准实验,验证了插值方法的有效性。

Registration method based on mutual information is an important method in the field of medical image registration,and it is popular in medical image processing fields,which is accepted as the most robust and accurate. When calculating the mutual information between two images,the image pixel points need interpolation in space transformation. At present,the interpolation algorithm has been all confined in single traditional interpolation method.Due to the disadvantages of the low interpolation accuracy or speed in traditional interpolation method,in this paper,we introduce a new image interpolation method which based on the absolute error of the pixel intensity.The interpolation algorithm combines the advantages of nearest neighbor interpolation algorithm and bi - cubic interpolation algorithm, thus improved the speed and the accuracy of the registration.This interpolation method has been verified through the registration experiment on head image.