An Efficient Biometric Based Remote User Authentication Technique for Multi-server Environment

Abstract

With the increasing demand of remote user authentication process for accessing many Internet based applications, biometrics based authentication mechanisms are highly being adopted. But using biometrics for these authentication mechanisms the features need to be captured and database need to be stored somewhere. Storing such a huge amount of biometric data itself has many security and privacy issues. Like revoking compromised biometric template is impossible. So template security is very important here. But due to low repeatability of biometrics, it is a challenging task to ensure template security while achieving high recognition performance of valid user. Hence, here we have proposed a keystroke based multi-server authentication mechanism which uses bio-hash template security technique. The bio-hash transform a biometric trait in an invariant feature which helps to increase the recognition performance. A comparison of normal keystroke dynamics and bio-hashed keystroke dynamics authentication performance has been done here. The bio-hashed keystroke shows better ERR of 0.15% compared to plain keystroke with 0.231% ERR.