A directional coding (DC) method is proposed to extract rotation invariant features for texture classification. DC uses four orientations in 3 × 3 neighborhood pixel. For each orientation, the rank order of the central gray level pixel is calculated. The four ranks are used to get 15 codes. The codes are combined with the information of the central pixel to extract 30 rotation invariant features. For a multi-resolution study, DC is calculated by altering the window size around a central pixel. The number of samples is restricted to eight neighbors by local averaging. Therefore, in each single scale DC histogram, the number of bins is kept small and constant. Outex, CUReT and KTH TIPS2 databases are used to evaluate and compare the proposed method against some state of the art local binary techniques and other texture analysis methods. The results obtained suggest that the proposed DC method outperforms other methods making it attractive for use in computer vision problems.
scite is a Brooklyn-based organization that helps researchers better discover and understand research articles through Smart Citations–citations that display the context of the citation and describe whether the article provides supporting or contrasting evidence. scite is used by students and researchers from around the world and is funded in part by the National Science Foundation and the National Institute on Drug Abuse of the National Institutes of Health.