A Novel Fingerprint Recognition and Verification System Using Swish Activation Based Gated Recurrent Unit and Optimal Feature Selection Mechanism

Abstract

Using fingerprints in biometric systems is a rapidly expanding and pervasive field. The advancement of fingerprint identification as a computer technology for applications is directly linked to the latest developments in computer science. A kind of fingerprint identification algorithm has been made possible by artificial intelligence technology; particularly imaging technology based on deep learning. This paper proposes a novel fingerprint recognition and verification system using a Swish activation-based gated recurrent unit (SWAGRU) with an efficient feature selection mechanism. The system mainly includes four phases: preprocessing, feature extraction, feature selection, and fingerprint recognition. To begin, the fingerprint samples are collected from the publicly available FVC2004 database. After that, Gaussian filtering is applied to the collected dataset to suppress the noise. Then, the feature extraction is carried out with the help of Self-Attention-Based Visual Geometry Group-16 (SAVGG16), and from that, the optimal features are selected based on Cuckoo Search Optimization (CSO). Finally, the fingerprint recognition and verification are done using SWAGRU. The experimental results showed that the system outperformed existing methods in recognition performance.