Gaussian Copula Multivariate Modeling for Image Texture Retrieval Using Wavelet Transforms

In the framework of texture image retrieval, a new family of stochastic multivariate modeling is proposed based on Gaussian Copula and wavelet decompositions. We take advantage of copula paradigm which makes it possibl e to separate dependency structure from marginal behavio r. We introduce two new multivariate models using respect ively generalized Gaussian and Weibull density. These mod els capture both the subband marginal distributions and the corelation between wavelet coefficients. We derive, as a simil arity measure, a closed form solution of the Jeffrey divergence between Gaussian Copula based multivariate models. Experimental resu lts on the well-known databases show significant improvements in retrieval rates using the proposed method compared to the bes t known state-of-the-art approaches.