A Novel Approach to Fingerprint Biometric-Based Cryptographic Key Generation and its Applications to Storage Security

Existing biometric-based security mechanisms to ensure storage security follow Biometric Enrolment, Key Binding, Secure Sketch, Fuzzy vault or template store along with key of a user. Also, they use threshold based comparison or error calculation to authenticate the user. The storage of biometric data or key put the system under threats. Further, the user verification mechanism may not be accurate as the threshold selection is challenging. To alleviate this problem, we have proposed a novel approach to storage security. In our proposed approach, we extract biometric based statistical features to generate codeword of a user. To generate a codeword, we use Reed-Solomon encoding (RS). Later, this codeword will be used to generate a key. Prio to decryption, we authenticate user using SVM Ranking mechanism without threshold value. The major contributions include generation of unique and strong biocrypto keys from users' biometric data, Reed-Solomon encoding has been used to maintain the codeword and generation of key and SVM based ranking mechanism is used for the user verification where the storage of neither templates nor keys is required.