Real Time Face Recognition for Mobile Application Based on Mobilenetv2

Abstract

Real-time facial recognition is one of the technologies with significant applications in a variety of contexts, including supporting the process of employee attendance. Attendance is a crucial aspect of company administration that influences productivity and operational effectiveness. Traditional attendance mechanisms are susceptible to fraud and errors, so businesses must adopt digital attendance solutions. Face recognition with landmark-based anti-spoofing using MobileNetV2 on mobile devices is intended to be an innovative solution for attendance management. This system employs CNN with MobileNetV2 architecture to detect and identify employee faces in real time. MobileNetV2 is advantageous because it makes efficient use of mobile device resources without sacrificing precision. The research results demonstrate the extraction of eye and lip landmark points with the Blazeface model integrated in the Dart programming language using the Flutter framework. The implementation of the mobile system yields an application called FaceON that can aid in the prevention of potential fraud by employing anti-spoofing techniques. Before a visage can be verified, users must overcome obstacles such as winks and smiles. The contribution of this study is that this system is a dependable and innovative solution for employee attendance management in the digital age