Motion-Sequence Authentication System: Guard for Smart Phones

EasyChair Preprint 7182, version 2

Abstract

In recent years, the mobile privacy protection is becoming increasingly critical due to the popularity of smartphones. The owner needs a “second line of guard” (user behavior authentication) when unlocking methods are attacked. The traditional user authentication approaches either face smudge attacks or can only work on dedicated devices. In this paper, we present a Motion-sequence Authentication System (MAS), an accurate and robust security authentication system that is not limited to expensive phones. MAS distinguishes user categories according to the unique characteristics of different user motion sequences. It is a rapid, non-contact, unobtrusive method of user authentication without predefined motions. MAS exploits Markov model to track the behavior of smartphone users, it can achieve real-time user authentication by utilizing the user’s short-term interaction with the smartphone. Our experiments in multiple real environments show that MAS can achieve higher than 94% accuracy for authenticating user motion sequence, which fills the gap with motion sequence recognition and provides a way of thinking for the development of human-computer interaction and information security.

Keyphrases: Human Computer Interaction, Markov model, Multi-scenario security authentication, behavior recognition

@booklet{EasyChair:7182,
  author    = {Yuzheng Dong and Yanfeng Zhao and Ziyue Wang and Juan He and Liubin Zhu},
  title     = {Motion-Sequence Authentication System: Guard for Smart Phones},
  howpublished = {EasyChair Preprint 7182},
  year      = {EasyChair, 2022}}