THE DEVELOPMENT OF INTELLIGENT BIOMETRIC IDENTIFICATION SYSTEM USER

Author(s):  S.N. Devitsyna, candidate of Sciences, associate Professor, Sevastopol State University, Sevastopol, Russia, Sevastopol

T.А. Eletskaya, Sevastopol State University, Sevastopol, Russia

T.N. Balabanova, candidate of Sciences, Belgorod State National Research University, Belgorod, Russia, sozonova@bsu.edu.ru

N.N Gakhova, candidate of Sciences, associate Professor, Belgorod National Research University, Belgorod, Russia, sozonova@bsu.edu.ru

Issue:  Volume 46, № 1

Rubric:  Accounting policies and measurement statistics

Annotation:  Biometrical methods are the effective tool for people identification and are used to verify the person, protect against unauthorized access, control the employee attendance, search for criminals, identify terrorists. The article provides an overview of modern biometric methods of identification, explain their advantages and disadvantages, as well as examples of their use in various fields. The analysis of statistical data and a comparison of the accuracy of different methods of biometric identification indicators according to FRR (False Rejection Rate) and FAR (False Acceptance Rate) which are the measures of reliability identification is also done. As a result, the face image was chosen as the identifier. To improve the accuracy of identification, the neural networks were used. An example of software implementation of an intelligent system of user identification by foto is given. The main difference from existing solutions is the simultaneous use of several machine learning tools during the images preparation and processing. The developed solution uses two neural networks for different tasks, which makes the identification accuracy higher in comparison with competitor solutions on the same type of biometrics.

Keywords:  information technology, information security, artificial intelligence, biometrics, identification, biometric technologies, pattern recognition

Full text (PDF):  Download

Downloads count:  303