Multimodal biometric cryptosystem based on fusion of wavelet and curvelet features in robust security application

Abstract

The authentication system used at the automated teller machines (ATM) is a unique personal identification number (PIN). This PIN can be easily tapped and misused. In this paper we propose a method in which the PIN is replaced by the biometrics of the individual to have a more secure transaction. A complete hardware system is designed to capture the biometric traits such as face, fingerprint and palm vein. The captured images are pre-processed and then features are extracted which are then fused at feature level. Cryptography is applied on the fused feature vector. Matching is performed using Euclidean distance at the server end. Palm vein is particularly chosen as a biometric trait along with widely used face and fingerprint because it is unique and is impossible to forge the vein pattern of an individual. Curvelet and Wavelet Transforms are used for the feature extraction. Experimental results indicate a good level of security and recognition rate of 91% and 89% is achieved on our own generated database. The results are promising when compared with other existing similar techniques.