Bimodal Eye Imaging System for Biometric and Medical Applications

article in a collection out of WoS and Scopus

English

The proposed contribution presents a new bimodal imaging system for simultaneous acquisition of the human iris and retina. The system is intended to use in high security applications for biometric human recognition (high-secured access) as well as in medical applications for possible use in ophthalmology. Biometric identification systems have become a common part of security applications. Iris or retina recognition is not a new idea, but if we consider these two as a multimodal system, it gives us a totally different view to the security and reliability of a respective biometric system. Common biometric systems in real-world applications are usually unimodal and thus can handle only one single biometric source of information (i.e. biometric characteristic) - usually a pattern of the human eye iris. Due to the fine and unique texture of iris, the probability of having the same iris texture is around 1 in 1078 [1], which ensures sufficient coverage of population. However, the fact the iris is located visibly on anterior segment of the eye and thus it is possible to take an iris photo from a distance of some meters, the risk of the iris pattern will be copied and subsequently counterfeited is relatively high. On the other side, the retina recognition is not currently used in practice at all. This is caused by several factors, e.g. complicated optical system, price, and low user-friendliness. Retinal recognition has clear advantages in uniqueness, number of features compared to other biometrics, and also that it is the only place in the human body with the possibility to observe the blood vessels non-invasively [2]. Hence, the retinal images can be utilized not only for biometrics, but also for medical purposes as a source of information for diagnosis, e.g. of various cardiovascular diseases, glaucoma, macular degeneration, diabetic retinopathy, and other disease that can be detected from retinal images. Unimodal biometric systems are generally susceptible to variety of issues such as intra-class variations, inter-class similarities, noisy data, non-universality, and spoofing. Hence, by combining iris and retina into one unique solution, it is possible to get very robust biometric recognition device. The proposed bimodal imaging system is designed as a part of the research project at the Brno University of Technology. The device allows simultaneous acquisition of iris and retina. Integrated illumination sources operate at both visible and near-infrared light. The whole optical system is mounted on an adjustable 3D platform which is controlled by a feedback from acquired images. Thanks to the 3D movable platform, we assume that the proposed biometric system will be more user friendly and in addition, the time of scanning will be much less than in case of other systems

eye imaging, iris, retina, biometrics, human recognition, security

KOLÁŘ, R.; SEMERÁD, L.; DRAHANSKÝ, M.; ODSTRČILÍK, J.; HÁJEK, J.; BOROVSKÝ, J.

2015

20. 5. 2015

University of Defence in Brno

Brno

978-80-7231-997-8

Security and Protection of Information 2015

83

95

13