Face Recognition Based on Improved Neighborhood Maximum margin
Youhu Rong, Kezheng Lin, Di Wu, Liangwei Zhuang · Advanced science and technology letters · 2014
In this paper, we proposed face recognition based on improved neighborhood maximum margin. The algorithm builds the k-nearest neighbor graph between the data points. Then assigns completely distinct weights for the interclass and intraclass neighbors of a point by fully considering the class information, and the algorithm formula ask the discriminant vector in range space of locality preserving between-class scatter and range space of locality preserving within-class scatter. At last, it finds a liner mapping by maximizing the margin between the interclass and intraclass neighbors of all points, to improve the classification performance in the new subspace. The experiment result on the UMIST face database shows that the method is feasible.