Graph Convolutional Network (GCN) Approach for Fingerprint Security
Hedi Boukamcha, Olfa Ben Ahmed · 2024
Cybersecurity has become paramount in our digital world. Fingerprint recognition offers a promising solution for biometric authentication due to its simplicity and reliability. However, traditional systems based on CNNs have limitations in modeling the relationships between fingerprint minutiae. To overcome these limitations, we propose a hybrid approach combining CNNs and GCNs. CNNs enable the extraction of precise local features from fingerprints, while GCNs model the spatial relationships between minutiae for better global representation.