Biometric-Based Key Handling Using Variation to Scale Invariant Feature Transform amid COVID-19 Pandemic

The outbreak of COVID-19 pandemic has hardened the biometric recognition, authentication, and cryptographic process based on facial features due to masks coverage. The Biometric cryptosystem (BCS) consisting of touch sensing biometric, such as fingerprint, palmprint, vein, is no longer in use in this scenario. In this context, other key-based biometric cryptosystems based on facial biometric may underperform. Thus, there is a need to simulate existing and implement novel biometric cryptosystems based on other biometric cues such as Ear, Face, Iris, Gait. This paper introduces a novel method for biometric key generation based on Scale Invariant Feature Transform (SIFT). The obtained key is used to encrypt/decrypt the data. At enrollment stage, the biometric key is obtained which subsequently encrypts the data and authentication phase decrypts the secret upon successful biometric match. The proposed method is compared with the state-of-art methods on three datasets namely AMI (ear), FACES (face), and UBIPr (iris). The metrics used to evaluate the performance of proposed method include: NPCR, UACI, PSNR, Correlation, MSE, MAE, SSIM, NRMSE.