Discriminative Robust Local Binary Pattern based Edge Texture Features for Object Recognition

Local Binary Pattern proves to be the most popular texture classification feature. The proposed system provides edge texture features, Discriminative robust Local Binary Pattern for the recognition. The algorithm used retains the contrast information and solves the issues of Local Binary pattern, Local Ternary Pattern and Robust Local Binary Pattern for proper representation. K-Nearest Neighbor classification and Surf matching techniques are used for classification and matching. The edge texture features obtained from the input image are stored and the image is retrieved based on the features extracted for the object of user?s interest. The new features are found robust to the image variations that are caused due to intensity inversion and are discriminative to the image structures within the histogram block. Local Binary Pattern is robust to the illumination and contrast variations. The proposed features also tend retain contrast information that is necessary for proper representation of the object contours.