Image fusion based on Immune Clone Selection and wavelet GΓD model

The computation of relative Saliency in fused approximate images´ local area is the key of obtaining efficient image fusion coefficients. According to the statistical characteristic of wavelet approximate coefficients in fused image, wavelet Generalized Gamma Distribution (G Γ D) model is built for capturing the distribution of approximate coefficients in local area. GΓD model´s parameters are estimated by Immune Clone Selection (ICS) algorithm derived from Evolutionary Computation thinking. The estimated parameters are applied in a new fusion strategy for obtaining the final fusion coefficients. Results clearly demonstrate the superiority of this new approach when compared to the methods such as traditional contrast pyramid (CP) and wavelet fusion methods in terms of image information entropy (IE) values, average gradient (AG) values and standard difference (STD) values.