Biometric Identification System’s Performance Enhancement by Improving Registration Progress

Abstract

Recently, biometric recognition techniques have become more and more important in security defense industry, among which stands out finger-vein identification technique, with distinctive advantages on accuracy, convenience, sanitation, safety, etc. Encouraged by its great features, we develop a series of finger-vein identification algorithms and apply them in a practical application system --- the Peking University Exercise Attendance System (PUEAS), which is based on finger-vein recognition technique. The system has been running for more than three years till now, accumulating more than 20,000 registered users, 900,000 finger-vein templates and 1,400,000 matching records. However, when we focus on how to make further improvement on the system, we find that the quality of the registration process plays a key role in determining the performance of the whole system. After discussing on some essential issues of the registration process, we conduct corresponding improvement measures to eliminate their influence. The experiment results well demonstrate enhancement of performance of the whole PUEAS.