Fusion of Multifocus Noisy Images using Contourlet Transform

Image fusion is one of the important application areas of image processing. It is used to obtain a single composite image with complementary features of various types of images. Presence of noise degrades the quality of an image and its subsequent fusion results, if it is not handled properly. Therefore, in this paper we have proposed an algorithm that combines denoising with fusion process. The proposed algorithm uses level dependent threshold that changes according to the characteristics of coefficients. The proposed method is compared with stationary wavelet transform and dual tree complex wavelet transform based fusion methods. The visual results show that the proposed algorithm gives better results than other methods. The performance of the algorithm is also tested using four quality metrics (peak signal to noise ratio, entropy, standard deviation and edge strength). Further, with the help of experiments, we have also established the fact that denoising before performing fusion gives effectively better results.