Abstract
As an important area in computer vision, textural image classification has been intensely investigated in the last decades. Among all existing methods, a recently developed image descriptor - local binary pattern (LBP) - has received tremendous attention because of its simplicity and robustness for representing textures. However, the selection of local binary patterns has been difficult, when the number of the patterns is very large and not all of them are discriminative in terms of representing textures. In this paper, we develop a new LBP operator (LBP_ex) for discriminative local binary patterns selection, and propose a directional Gaussian filter-based LBP_ex descriptor for textural image classification through two major steps: 1) using a bank of directional Gaussian filters to retrieve the anisotropic information in the textural images; 2) combining the LBP_ex histograms calculated from both original images and filtered images to form feature vectors that represent isotropic and anisotropic properties of the texture images in attempt to further improve the classification accuracy. We experimentally evaluate the performance of the method through comparing with four existing state-of-the-art LBP algorithms on the same database OUTex, and the results demonstrate that the features represented by the new LBP_ex descriptor are more discriminative, leading to a superior performance.
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Disclaimer: All third-party content on this website/platform is and will remain the property of their respective owners and is provided on "as is" basis without any warranties, express or implied. Use of third-party content does not indicate any affiliation, sponsorship with or endorsement by them. Any references to third-party content is to identify the corresponding services and shall be considered fair use under The CopyrightLaw.