Article

Article title USER IDENTIFICATION BASED ON VIBROACOUSTIC NOISES ORIGINATED FROM ARBITRARY TEXT TYPING
Authors V.M. Fedorov, E.M. Panchenko, D.P. Rublev
Section SECTION V. APPLICATION INFORMATION SECURITY ISSUES
Month, Year 12, 2013 @en
Index UDC 004.056:061.68
DOI
Abstract In presented work the user identification task based on vibroacoustic signal originated from typed arbitrary text is reviewed. To increase identification accuracy the method of keypresses/releases pauses removal was developed. A training sets consisting of vibroacoustics signals of keypresses/releases and vibroacoustics noises in pauses has been developed. Obtained sets were used as training data for neural net. Pauses detection was done in running window with 1024 samples length overlapped by Ѕ of its length. If window fragment was detected as belonging to pause segment, this fragment was removed. Features for detection user identity were considered and stable set of features was selected, namely: Fourier transform coefficients, cepstral coefficients and linear prediction coefficients. User identification methods based on statistical parameters, mixed Gaussian models and neural nets were reviewed. Feasibility of user identification based on analysis of vibroacoustics noises originating from typing was shown, probability of successful user identification on created base of vibroacoustics signals originated by several users was estimated.

Download PDF

Keywords Vibroacoustic signal; discrete Fourier transform; cepstral coefficients; neural nets; linear prediction coefficients; Gaussian mixed models; identification; access control.
References 1. Ilonen J. Keystroke Dynamics // Lappeenranta University of Technology. 2008.
2. Фёдоров В.М., Рублёв Д.П. Методы предварительной обработки виброакустических сигналов от клавиатуры возникающих при наборе текста // Информационное противодействие угрозам терроризма. – 2012. – № 18. – С. 172-175.
3. Фёдоров В.М., Рублёв Д.П. Фильтрация виброакустических сигналов от клавиатуры и манипулятора мышь, возникающих при работе оператора // Информационное противодействие угрозам терроризма. – 2012. – № 19. – С. 160-162.
4. Федоров В.М., Рублев Д.П. Обработка виброакустических шумов, возникающих при работе пользователя с клавиатурой // Известия ЮФУ. Технические науки. – 2012. – № 12 (137). – С. 75-81.
5. Бабенко Л.К., Макаревич О.Б., Федоров В.М., Юрков П.Ю. Голосовая текстонезависимая система аутентификации (идентификации) пользователя // Известия высших учебных
заведений. Радиоэлектроника. – 2004. – Т. 47, № 3. – С. 66-70.

Comments are closed.