Face recognition system using Multilinear Principal Component Analysis and Locality Preserving Projection

Face recognition technology has evolved as an enchanting solution to perform identification and the verification of identity claims. By advancing the feature extraction methods and dimensionality reduction techniques in the pattern recognition applications, number of facial recognition systems has been produced with distinctive degrees of success. In this paper, we have presented the biometric face recognition approach based on Multilinear Principal Component Analysis (MPCA) and Locality Preserving Projection (LPP) which enhance performance of face recognition. The methodology of the approach consists of face image preprocessing, dimensionality reduction using MPCA, feature Extraction using LPP and face recognition using L2 similarity distance measure. The proposed approach is validated with FERET and AT&T database of faces and compared with the existing MPCA and LDA approach in performance. Experimental results show the effectiveness of the proposed approach for face recognition with good recognition accuracy.