A Novel Bio-Inspired Texture Descriptor based on Biodiversity and Taxonomic Measures

Texture can be defined as the change of image intensity that forms repetitive\npatterns, resulting from physical properties of the object's roughness or\ndifferences in a reflection on the surface. Considering that texture forms a\ncomplex system of patterns in a non-deterministic way, biodiversity concepts\ncan help texture characterization in images. This paper proposes a novel\napproach capable of quantifying such a complex system of diverse patterns\nthrough species diversity and richness and taxonomic distinctiveness. The\nproposed approach considers each image channel as a species ecosystem and\ncomputes species diversity and richness measures as well as taxonomic measures\nto describe the texture. The proposed approach takes advantage of ecological\npatterns' invariance characteristics to build a permutation, rotation, and\ntranslation invariant descriptor. Experimental results on three datasets of\nnatural texture images and two datasets of histopathological images have shown\nthat the proposed texture descriptor has advantages over several texture\ndescriptors and deep methods.\n