Logarithmic high dimensional model representation in image processing

Abstract

Logarithmic HDMR is an HDMR based method and it allows us to reconstruct an analytical structure for a given multivariate analytical function in terms of less variate components. In this study, we develop a method to model multivariate data through Logarithmic HDMR and we use this method to make an approximation to the given original image. That is, we deal with the reconstruction of an image by using Logarithmic HDMR philosophy and we reach very promising results.