Hybrid Approach to Face Recognition System using PCA & LDA in Border Control

Abstract

In this paper, we have proposed a hybrid approach to face recognition using PCA and LDA in border control like time-critical application. Duplicate passports, unauthorized VISA, fake identities, and border criminals have dramatically increased within the last few years in Sri Lanka due to a lack of proper identification system at the borders. Even though in the traditional VISA granting process biometrics are sent to borders and final destinations such as immigration counters, there are no proper face recognition measures at airports, to prove that it is the same person who comes to apply VISA that is identified at the airport. The proposed system uses Principal Component Analysis (PCA) and Linear Discriminant Analysis (LDA) to compare updated face biometrics and the physical appearance at the airport in a novel hybrid way. After the face recognition is done by a hybrid solution with a confidence level, the candidate will proceed to the manual process as usual. The proposed system speeds up the traditional process, and increasing the accuracy of identification and smooth adaptation to the traditional method can be identified as the main benefits of the system. Face biometrics is the main ingredient of the proposed system. The system has been analysed with the traditional model and evaluated with authentic biometric sample and identified with 98% accuracy in face recognition with less average time. A nearest mean hybrid approach in the time-critical application can be identified as the novelty of the proposed system.