Texture analysis by fractal descriptors over the wavelet domain using a best basis decomposition

This work proposes the development and study of a novel set of fractal descriptors for texture analysis. These descriptors are obtained by exploring the fractal-like relation among the coefficients and magnitudes of a particular type of wavelet decomposition, to know, the best basis selection. The proposed method is tested in the classification of three sets of textures from the literature: Brodatz, Vistex and USPTex. The method is also applied to a challenging real-world problem, which is the identification of species of plants from the Brazilian flora. The results are compared with other classical and state-of-the-art texture descriptors and demonstrate the efficiency of the proposed technique in this task.