A structure-preserved local matching approach for face recognition

In this paper, a novel local matching method called structure-preserved projections (SPP) is proposed for face recognition. Unlike most existing local matching methods which neglect the interactions of different sub-pattern sets during feature extraction, i.e., they assume different sub-pattern sets are independent; SPP takes the holistic context of the face into account and can preserve the configural structure of each face image in subspace. Moreover, the intrinsic manifold structure of the sub-pattern sets can also be preserved in our method. With SPP, all sub-patterns partitioned from the original face images are trained to obtain a unified subspace, in which recognition can be performed. The efficiency of the proposed algorithm is demonstrated by extensive experiments on three standard face databases (Yale, Extended YaleB and PIE). Experimental results show that SPP outperforms other holistic and local matching methods.

Research highlights
► Compared with most existing local matching methods (such as SpPCA, Aw-SpPCA and SpNMF), a main advantage of SPP is that it takes the configural information of sub-patterns into consideration and can preserve the configural structure of each input face image during feature extraction. In our study, the configurations of each face image are characterized by linear coefficients which can reconstruct the given sub-pattern by other sub-patterns from the same face image.
► The proposed SPP uses the locality preserving technique to preserve the local neighborhood information of each sub-pattern set. Therefore, the essential manifold structure of the sub-pattern sets can also be preserved by our method.
► In order to utilize the label information of the training samples for recognition, we extended the proposed algorithm to a supervised mode. From the experimental results, it can be found that the supervised information can improve the performance of the proposed method, especially when the sub-pattern size is large.