Article

Article title PARAMETRIC METHODS OF RECOGNITION FORM A DYNAMIC BIOMETRICS
Authors Yu.A. Bryukhomitsky
Section SECTION V. BIOMETRIC AND IMMUNOLOGICALMETHODS OF PROTECTION OF THE INFORMATION
Month, Year 12, 2011 @en
Index UDC 004.065
DOI
Abstract We consider three parametric method of pattern recognition biometrics for use in dynamic biometric authentication systems by individual handwriting and keyboard writing style. The first method is based on the computation of the quadratic discriminant function coefficients are determined by the characteristics of the distribution of biometrics "own" the user. In the second method discriminant function is presented and evaluated as a multivariate density function biometric user. The third method uses the canonical form of the quadratic discriminant function. The proposed methods have a high recognition accuracy and ease of learning.

Download PDF

Keywords Dynamic biometric authentication system; a multivariate normal distribution of biometric parameters; parametric methods of pattern recognition; discriminant function; training; verification.
References 1. Болл Р.М., Коннел Дж.Х., Панканти Ш., Рахта Н.К., Сеньор Э.О. Руководство по биометрии. – М.: Техносфера, 2007. – 386 с.
2. Нильсон Н. Обучающиеся машины: Пер. с англ. – М.: Мир, 1967. – 180 с.
3. Беклемишев Д.В. Курс аналитической геометрии и линейной алгебры.  М.: Наука, 1971.  328 с.
4. Брюхомицкий Ю.А., Казарин М.Н. Учебные биометрические системы контроля доступа по рукописному и клавиатурному почеркам // Сборник научных трудов XIII Всероссийской научной конференции «Проблемы информационной безопасности в системе высшей школы». – М.: Изд-во МИФИ, 2006. – С. 33-34.

Comments are closed.