A novel image fusion algorithm based on regional Kullback-Leibler entropy analysis and nonsubsampled contourlet
transform is proposed in this paper. The equation of Kullback-Leibler entropy is modified at first, and then the modified
Kullback-Leibler entropy of the corresponding area of the two source image is calculated. The result of the
Kullback-Leibler entropy is clustered to three classes. According to the result of the clustering, different fusion strategies
are selected for low frequency subband coefficients. High frequency coefficients are fused using a "local feature-based"
rule. Then the fused coefficients are reconstructed to obtain the fused image. Experimental results showed that the
proposed algorithm not only improved the visual effect, but also enhanced the contrast and information entropy.