关键词: 人脸识别, 近邻保持嵌入, 最大散度差准则

Abstract:

Neighborhood preserving embedding (NPE) emphasizes the face sample manifold local structure, without taking into account the sample class discriminant information and the small sample problem. The algorithm is based on NPE and the maximum scatter difference criterion (MSDC) is introduced to its objective function. After being embedded into a low dimensional subspace, the samples of the same class maintain their intrinsic neighbor relations while the samples of the different classes are far from each other. And then the most discriminative feature is extracted. Experiments in the AT & T face database show that the highest recognition rate of the algorithm has increased by 15.35%, 6.47%, and 6.94%, respectively, comparied with the PCA, LDA and NPE algorithm. Comparative experiments in the Yale face database show that the highest recognition rate of the algorithm has increased by 20.27%, 5.63%, and 2.27%, respectively.

Key words: face recognition, neighborhood preserving embedding, maximum scatter difference criterion