Securitas: user identification through RGB-NIR camera pair on mobile devices

Abstract

Today mobile devices are equipped with numerous sensors and new ones are being added. In this paper, we propose a method to utilize a new sensor to provide a more secure identification system named Securitas for mobile device users. Securitas is a user identification system through the use of RGB-NIR camera pairs. The system extracts and analyzes geometrical features from a human hand to identify the user for unlocking devices and accessing personal data. Utilizing both RGB and the NIR cameras for real skin detection, it can effectively prevent an impostor from gaining access by using a fake hand photograph of a valid registered user without limitations of contrast, color, and background. Comparing to existing techniques, Securitas demonstrates that by leveraging the sensors on the mobile devices, a user can have a more secure identification mechanism by simply taking a photograph of his hand. Through proof of concept of implementation, our system demonstrates the ability to distinguish users with more than 94% accuracy.