Estimating mutual information of image by fast Gaussian transform
应用快速高斯变换估计图像互信息

A new fast algorithm was presented to accelerate the computation of mutual information of images based on kernel density estimate. This algorithm adopted gauss function as kernel function, and then used fast Gaussian transform to reduce time complexity and improved the fast Gaussian transform by k-center clustering. The new algorithm can evaluate the mutual information in linear time complexity and experiments show it is very efficient.