User re-authentication via mouse movements

摘要: We present an approach to user re-authentication based on the data collected from computer's mouse device. Our underlying hypothesis is that one can successfully model behavior basis of user-invoked movements. implemented system raises alarm when current X, deviates sufficiently learned "normal" X. apply a supervised learning method discriminate among k users. empirical results for eleven users show we differentiate these individuals their movement with false positive rate 0.43% and negative 1.75%. Nevertheless, point out analyzing movements alone not sufficient stand-alone system.