Phone Pick-up Authentication : A Gesture-Based Smartphone Authentication Mechanism

Passcodes, fingerprints and now facial scans, smart-phone authentication is becoming ever faster and faster and easier to use. Providing one time authentication mechanism that operate quickly an integrate ever more seamlessly into their customers' lives, is a concern for all mobile companies. These older methods are however more liable to spoofing and smudging attacks and as a result there is a call for new and improved methods. Gesture authentication looks to be a promising area of research but most gesture based schemes requires a specific input from the user, detracting from the simplicity of verifying oneself. We developed a novel gesture based authentication mechanism capable of recognizing the user based on the motion used to remove a phone from their pocket. Through this we are able to provide a high degree of security, but maintaining the user-friendliness of a fast user authentication approach, like fingerprint scanning. Our method also circumvents issues like wearing gloves or face masks. To prove the viability we collected over 500 samples from 10 users taking a mobile phone out of their pocket. Our models were able to achieve an authentication accuracy of over 93% only using 15 valid samples for training and a identification accuracy of 94%.