Conventional Local Binary Pattern (LBP) methods follow the patterns whose rotations are lesser than two or certain limited numbers are called rotation invariant binary patterns. In the conventional rotational-invariant encoding method has disadvantage due to neglecting information of the some patterns by its process of encoding. It ignores the patterns when their spatial transition is greater than two for maintaining the rotation-invariant nature. But these disregarded patterns will plays crucial role and have very much more discriminative power. Here, the present study proposing a novel model called OLBP by changing (sorting) the order of consecutive binary patterns without disturbing the property of rotational invariance. The result observed by experiments indicates the proposed work shows better classification rate which is worked on the standard databases when compared to previous existing methods.
Muthevi, A. K., & Uppu, R. B. (2019). Ordered local binary pattern (OLBP) for classification of textures. International Journal of Recent Technology and Engineering, 7(5), 243–247.