Fingerprint-Based Asymmetric Bio-Cryptographic Key Generation Using Convolution Network

The asymmetric cryptography techniques use in the field of data security suffer from the challenge of maintaining the private key as secret. As a result, creating cryptographic keys utilizing an individual user's bio-metric traits is a viable option. Convolutional operations are an excellent method for extracting information from bio-metric images. In the proposed work, a asymmetric bio-key generation process is formulated using convolution network for feature extraction with polynomial solver and prime factorization. A large number of asymmetric bio-key can be generated from the proposed method. Three publicly available data set called FVC2000, FVC2002 and FVC2004 are used for the experiment. Relevant security analysis is performed to evaluate the proposed algorithm.