Employee Fingerprint Identification Management System Based on Bidirectional ResNext Network with Triplet Loss

Abstract

With the increasing digitization and modernization of enterprise management, employee fingerprint identification management systems have become a key tool to ensure the safety and efficiency of the office space. Such a system not only accurately verifies the identity of employees but can also be used for attendance, access control and other related applications. However, improving the accuracy of fingerprint recognition remains a challenge. In this paper, we propose a fingerprint identification method based on a bidirectional ResNext network and combined with triplet loss. Compared with the traditional ResNext network, our method captures richer fingerprint features through bidirectional structure and triplet loss, thereby achieving higher matching accuracy. We conducted extensive experiments on the public FVC (Fingerprint Verification Competition) dataset and our own collected dataset CSFI (Company staff fingerprint identification). Experimental results show that compared with the original ResNext network, our method achieves approximately 3% and 5% performance improvements on the two data sets respectively. These results verify the effectiveness of our proposed method on fingerprint identification tasks and provide strong technical support for the further development of employee fingerprint identification management systems.

Topics & Concepts