期刊名称：International Journal of Advanced Research in Computer Engineering & Technology (IJARCET)
印刷版ISSN：2278-1323
出版年度：2014
卷号：3
期号：5
页码：1711-1715
出版社：Shri Pannalal Research Institute of Technolgy
摘要：Mouse gesture dynamics is a special type of behavioral biometric authentication technique. This approach extracts and analyses the mouse movement characteristics by using a computer mouse as a pointing device. In general, there are two types of authentication i.e. continual authentication and static authentication. Continual authentication system has been widely used and evolution of such system is much faster when compared to static authentication based system. Signature drawn by the user as input for the static authentication and mouse lock method are the only two existing mouse based static authentication systems. In order to improve the efficiency of static authentication based system, we are introducing a new model to make the system with much higher efficiency and robust. In this new approach user draws, at long time, a few gestures from which the gestures are collected and analyzed for authentication purpose. Existing gestures based authentication systems uses other input devices such as stylus but we use mouse as our input device for capturing the gestures. The captured gestures are analyzed using a hidden markov model classifier. Mouse gesture dynamic system yields impressive FAR of 5.2% (False Acceptance Ratio) and FRR of 4.5 % (False Rejection Ratio) compared to the existing mouse based static authentication systems. This Improvement in both accuracy and validation compared to existing mouse dynamic approaches that could be considered adequate for static authentication. To the best our knowledge, our work is the first to present an accurate static authentication scheme based on mouse gesture dynamics.
