Adaptive Image Fusion Scheme Based on Contourlet Transform, Kernel PCA and Support Vector Machine

A novel image fusion algorithm based on the combination of contourlet transform, Kernel Principal Component Analysis (K-PCA), Support Vector Machine (SVM) and Mutual Information (MI) is proposed. Image representation is improved by utilizing the Contourlet transform properties, such as localization, multiresolution, directionality and anisotropy. K-PCA is used to extract features in the low frequency subband and SVM operates (in high frequency subbands to obtain a composite image with extended information. Moreover, Mutual Information (MI) is used to adjust the contribution of each source image in the final fused image. Performance of the fusion algorithm is evaluated by recently developed metric, Image Quality Index. (IQI). Experimental results show that the proposed method outperforms previous approaches both subjectively and quantitatively.