Fingerprint Recognition System based on Big Data and Multi-feature Fusion
Mikai Yang · 2020
This paper proposes a fingerprint recognition algorithm, which is based on deep learning and multi-feature fusion, is able to solve the shortcomings of traditional fingerprint recognition that purely relies on minutiae features for recognition, especially in the case of partially incomplete fingerprints or small-area fingerprints. For identification, while the user inserts a fingerprint in the terminal, the system will transmit raw data to the deep learning engine of the cloud in an anonymous and encrypted manner, which can quickly collect a large amount of fingerprint data in an aligned manner. Based on these data, parameters of the deep-learning model are adjusted incrementally, pursuing optimal accuracy of the system. Meanwhile, SaaS services are provided to other applications for fingerprint recognition.