Designing of a Biometric Multimodal Verification System Using Contactless Sensor Based on Information of the Texture of The Hand

One of the most impressive tools for evaluation of Cyber security is the usage of the secure verification system with high reliability. Biometric based verification systems have been able to overcome this problem to a large extent. In many cases, some consideration such as hardware cost, speed and accuracy make challenges to choose the architecture of verification system. In this paper, an ordinary sensor is used to increase the security and accuracy of the verification system and also decrease the hardware cost. After acquiring the hand image, it is divided into 6 segments and finally by using Arc-Gabor filter bank as a descriptor, unique features are extracted from images. Also, by extracting knowledge from the FAR and FRR curves in order to estimate verifiers’ weight and using weighted score summation for information fusion in the level of matching score, the verification system is designed. In this system, by utilizing one ordinary sensor and potential of multimodal architecture, its average of Equal Error Rate for COEP database is around 2.25% and also the average of elapsed time for verification is less than 0.19 seconds.