Color texture image retrieval based on Copula multivariate modeling in the Shearlet domain

Due to the increasing number of images on Internet, image retrieval framework is needed for image-based search. Shearlet transform a sparse, multiscale, and multidimensional representation used extract anisotropic features in processing applications such as fusion denoising. In this paper, we proposed color texture based domain modeling using Copula multivariate model. framework, Gaussian model dependencies between different sub-bands Non-Subsample Transform (NSST) non-Gaussian models are marginal coefficients. Six schemes NSST coefficients four types neighboring defined; moreover, Kullback–Leibler Divergence(KLD) close form calculated situations two functions order investigate similarities framework. The Jeffery divergence (JD) criterion, which symmetrical version KLD, investigating We have implemented our experiments benchmark datasets, results show superiority over existing state-of-the-art methods. addition, time also analyzed steps feature extraction similarity matching, shows that enjoys an appropriate time.