ANALYSIS OF THE IDENTIFICATION ERRORS AND WAYS OF ACCURACY IMPROVEMENT OF THE SYSTEMS OF BIOMETRICAL IDENTIFICATION
DOI:
https://doi.org/10.31649/2307-5392-2020-2-1-9Keywords:
biometrical identification, biometric profiles, recognition, vein print, biometric system, scanner, identification accuracy, biometricsAbstract
Identification and authentication provides safe protection of the personal data ,realizes reliable access to the objects or confidential information, results of scientific research, etc. Especially important becomes the combination of the passwords with the biometrical characteristics of a person, as a result , the reliability of the access system increases thousandfold. The paper considers basic properties and principles of the biometric identification systems, errors of the main methods of biometric identification have been analyzed and classified: Static methods ( fingerprint identification, iris identification, face or hand geometry, venous picture of the hand ,eye retina) and dynamic(voice identification, fingertrip identification, signature identification). Principle of hand vein pattern obtaining, which consists in IR absorption by blood hemoglobin has been substantiated, according to which the necessary vein picture is determined by the degree of the emission reflection and the level of the blood stream visualization. The comparative analysis of several methods of identification by the veinprints has been performed and the method with the best characteristics has been selected. The analysis of the control points of the vascular pattern (minutia) has been performed. For optimal data taking new technique of hand scanning, using blocking device is introduced, the given technique enables to reduce to minimum the veinprint shift of the investigated person and makes its comparison with the sample easier. The choice of the commercial sample of the vein reader – scanner, which can be quickly and qualitatively adapted in the general structure of the identification system is substantiated. For the skeletonization of the binary image the method of patterns is selected, this method enables to process the data rapidly and obtain the data matrix – coordinates of the minutia and their type. For the improvement of the functional model of the person identification process by the hand vein pattern the identification protocol was introduced to the structure of the identification process, the realization of this protocol stipulates adequate taking a decision “identified / unidentified”.
Downloads
-
PDF
Downloads: 137