Verification of pattern unlock and gait behavioural authentication through a machine learning approach

Abstract

PurposeThe existing authentication procedures (pin, pattern, password) are not very secure. Therefore, the Gait pattern authentication scheme is introduced to verify the own user. The current research proposes a running Gaussian grey wolf boosting (RGGWB) model to recognize the owner.Design/methodology/approachThe biometrics system plays an important role in smartphones in securing confidential data stored in them. Moreover, the authentication schemes such as passwords and patterns are widely used in smartphones.FindingsTo validate this research model, the unauthenticated user's Gait was trained and tested simultaneously with owner gaits. Furthermore, if the gait matches, the smartphone unlocks automatically; otherwise, it rejects it.Originality/valueFinally, the effectiveness of the proposed model is proved by attaining better accuracy and less error rate.