Unsupervised Gabor Filter-Bank Method for Characterization of the Self-Assembled Hexagonal Lattice

Abstract

In this paper a new robust and precise ordering criterion for the characterization of self-assembled hexagonal lattice like Anodic aluminum Oxide (AAO) has been proposed. In order to unveil the mechanism for the self-organization process and deposition techniques in AAO, it is necessary to be able to have a quantitative objective criterion to evaluate the amount of order through every SEM sample of a material. Most of methods in the literature are only able to characterize the extreme case of highly ordered or lowly ordered texture well. But the real challenge is in characterizing the order of medium-ordered texture which is the dual concept of near-regular texture analysis in image processing. Our method based on more advanced image processing techniques, Gabor filter-bank, are able to characterize medium-ordered AAO textures more precisely. Our idea is also able to define the domain's place of the AAO image.