Credential hardening by using touchstroke dynamics

• Main Authors: Teh, Pin Shen, Teoh, Andrew Beng Jin, Yue, Shigang
• Format: Article
• Language: English
• Published: Sunway University 2014
• Subjects: QA75 Electronic computers. Computer science
• Online Access: http://eprints.sunway.edu.my/404/1/SAJ_Vol%2011_35-54.pdf; http://eprints.sunway.edu.my/404/

Today, reliance on digital devices for daily routines has been shifted towards portable mobile devices. Therefore, the need for security enhancements within this platform is imminent. Numerous research works have been performed on strengthening password authentication by using keystroke dynamics biometrics, which involve computer keyboards and cellular phones as input devices. Nevertheless, experiments performed specifically on touch screen devices are relatively lacking. This paper describes a novel technique to strengthen security authentication systems on touch screen devices via a new sub variant behavioural biometrics called touchstroke dynamics. We capitalize on the high resolution timing latency and the pressure information on touch screen panel as feature data. Following this a light weight algorithm is introduced to calculate the similarity between feature vectors. In addition, a fusion approach is proposed to enhance the overall performance of the system to an equal error rate of 7.71% (short input) and 6.27% (long input).