Security Enhanced Rhythm Authentication using Relative Finger-Tip Positions

Abstract

The rhythm authentication algorithm uses the concept of a traditional keystroke authentication and a multitouch technology based on a touchable device. Three measurements are used in the rhythm authentication method which consists of holding times, latency times, and the number of fingers per beat. These measurements are compared with the user templates. The user is authenticated if a percentage of error is in a predetermined range. However, using only three measurements is not enough. If the attacker is familiar with the victim, the rhythm authentication can be attacked by shoulder surfing or eavesdropping. Additionally, only a percentage of similarity between the user template in database and user’s input for classifying is not a reliable algorithm. In this research, we propose a security-enhanced rhythm authentication using relative finger-tip positions and the KNN algorithm for classification to prevent shoulder-surfing attacks.