Fingerprint Recognition System based on Deep Learning Architecture

• A customized deep learning based fingerprint recognition system has been developed using the multitask deep convolutional neural network architecture to extract the fixed length representation from a high res- olution fingerprint image. • The recognition system calculates the cosine similarity score be- tween the two representations to generate a distribution of genuine and imposter scores.