Combining of feature extraction for real-time facial authentication system

The face is one of the biometric personal identification in access to services both private and public which uses authentication system. Naturally it is easy recognized from long distance, unfortunately its performance is still low compared to other biometric. To improve performance of face recognition there should be some of feature extraction techniques because this section is the basic process for authentication purposes by combining. This paper presents the combining of two feature extraction are principal component analysis (PCA) and linear discriminant analysis (LDA) using Euclidean distance classifications to improve the performance of face-based real time authentication system for users. Combining of feature extraction to calculate values of PCA and LDA so are acquired the matching scores. The matching scores are sums of both of feature extraction values. The combined extractor can improve distinctive values of face images. Distance measure to find the shortest distance between the template images with the test images using Euclidean distance. The system performance provide authentication of personal identification based on each face image of user. Results obtained from image of the 11 users show that combined extractor give the best performance compared to the performance of a single feature extraction. Proposed method provides better average performance than single extractor, it can solve low performance on facial authentication system. The real time system will be more flexible and compatible with the existing conditions of services so need better performance.