Energy Efficient Multifocus Image Fusion in DCT Domain

Abstract

Fusion of multifocus images in discrete cosine transform (DCT) domain has been extensively exploited in recent years due to low complexity. In particular, the fusion by DCT is useful in visual sensor networks (VSN), where the images have to be transferred in coded format. Many research works have been done by combining spatial domain methods in DCT domain. In the proposed work, the energy of correlation coefficient in DCT domain is chosen as fusion criteria. The method works by evaluating the focus measurement between the input images and Laplacian-based sharpened images in DCT domain. The results obtained by the proposed method are compared in terms of objective metrics and the results show that the proposed work avoids inappropriate block selection which exists in available methods.