Adaptive approach for Image Fusion using Curvelet Transform and Genetic Algorithm

Abstract

Although image fusion is a technique of merging two or more images that have consilient information to form a fused image which contains more accurate information of the image than any of the individual source images. In this paper, we proposed a multi-view and multi-modal Fusion, and Pixel level fusion approach. At first stage we perform feature extraction of image which plays a major role in the implementation of fusion approaches. Prior to the merging of images, salient features, present in all source images, are extracted using an appropriate feature extraction procedure. For the same we use transform domain texture feature Extraction (Curvelet) for better edge representation. After that fusion is performed on these extracted features vector by using genetic algorithm to get the more optimized combined image. Performance evaluation has been carried out of using the RMSE, PSNR and IQI. The results of the proposed method is compared with the existing techniques of image fusion using DWT. Experimental results shows that of curvelet transform and GA is better than DWT fusion method.