An approach to image fusion implementing feature level fusion on seismic attributes [articol]

Abstract

Image fusion is an important part of image processing, being performed on different levels of complexity. A feature, or attribute level fusion between two or more images, could be performed by combining the notions of image segmentation and mutual information in order to obtain the fused image. The first step of the proposed method is to decompose the input images into basic features using a connected-components algorithm, followed by the computation of the mutual information using one of the inputs as reference. The fusion method will be tested on synthetic images and further applied on seismic attributes.