Machine Learning Face Recognition Model for Employee Tracking and Attendance System

Abstract

An attendance system is widely implemented to monitor someone's presence in the office, schools, or events. Several technologies can be implemented as the attendance system. Face recognition is a natural, inexpensive, and easy way to be implemented as an attendance system. In this pandemic and post-pandemic era, face recognition can be the best alternative to be implemented as the attendance system. This research aims to propose a real-time attendance system using face recognition with consistently high accuracy. Moreover, the system is able to update the attendee's face periodically to tackle the changes in the face over time. The attendee can take their attendance using a camera. The camera captures the face and detects the face using the Multi-Task Cascaded Convolutional Neural Network (MTCNN). In addition, The Face Alignment N etwork (FAN) is applied to the image to extract the facial landmark in the image. The next step is to extract information from the face by using FaceN et. Finally, the face embedding extracted from Face Detection System is classified. The best classifier accuracy achieved by the model was 100% and 99.90% for training and validation, respectively.