BioSmart company operates in the market of biometric systems since 2006 and during this period it was able to establish itself as a reliable developer.
Within the confines of the largest international exhibition Securika 2019, dedicated to security technology, safety and fire protection equipment, issues of biometric identification, reliability and accuracy of biometric algorithms were discussed. As a result of the discussion a very important thought was expressed: any biometric identification system, however accurate it might be, would be useless if it is easily deceived. Techniques of cheating and countering this are reviewed in more detail below.
One of the key features of any biometric access control system is the ability to reliably detect fakes, by which malefactors try to gain unauthorized access. BioSmart system contains elements of a full-featured software and hardware complex ensuring the implementation of this requirement.
Infrared radiation (IR) is an integral component of the device for scanning the venous pattern of the palm. There is a definite dependence of the reflection, penetration and absorption coefficients of IR rays for various materials depending on the wavelength. It is this feature, which underlies the hardware of the complex for dummies detection. BioSmart system uses multispectral IR illumination, which makes it possible to produce a series of palm shots at different wavelengths. The subsequent analysis of the obtained data allows to qualitatively define to which group of materials the object presented to the BioSmart system may be classified and make the appropriate decision.
Besides, the system is additionally equipped with software protection against fakes. The image of the palm veins is carefully analyzed using especially developed classification algorithms. The algorithm accepts a digital image as input, extracts the necessary features from it and classifies the object in the image as a living palm or as a dummy. Software protection is developed using modern methods of machine learning, texture and morphological image analysis.
At the end of 2018, researchers from the Technical University of Berlin Jan Krissler and Julian Albrecht developed the way to deceive the method of biometric identification according to the drawing of blood vessels using a specially prepared and printed pattern that was waxed to imitate the skin. BioSmart developers constantly explore new techniques of possible cheating the system, develop solutions to identify fakes and, if necessary, improve the algorithms. The identification and development of ways to counteract the methods of cheating the identification system described above through the palm veins were implemented in the company 4 years ago, and then they were included at the hardware and software level in all commercially available BioSmart devices.