Local Binary Pattern for Binary Object Classification using Coordination Number (CN)* and Hu's Moments
Ratnesh Kumar, Kalyani K Mali · 2021
Hu's Moment corresponding to the Local Binary Pattern (LBP) features vector have been frequently used for binary objects pattern recognition and classification. Hu's Moments are used in a variety of applications due to its geometrically invariant features on object translation, rotation and scaling. In this paper, we have formulated the methodology of the Hu's Moment by using Local Binary Patterns. The minimum responding decimal values of the developed model for the Local Binary Pattern (LBP) corresponding to the Coordination Number (CN)* of each contour point of an object are used to compute Hu's seven moments. We have incorporated the local binary patterns corresponding to the coordination number of object contour points in a modified model for computing similarity between two binary objects. The experimental results conducted on Kimia-99 and MPEG-7 CE datasets demonstrate that our approach shows better results compared to the other existing algorithms.