FACIAL RECOGNITION ATTENDANCE SYSTEM

Abstract

Colleges have historically faced a great deal of difficulty with student attendance, necessitating a large time and effort investment from staff in manual tracking. Even though they are in place, the existing biometric attendance systems are not entirely automated, which causes delays in processing fingerprints, maintenance issues, and inefficiencies in time. Given that almost everyone has a smartphone and is continuously online in this day and age, a more simplified method is necessary. This study suggests using sophisticated object identification algorithms to check attendance using faculty members' smartphones. Because of its effectiveness in face detection and the addition of Microsoft Azure's face API for database recognition, YOLO V3 (You Only Look Once) is the preferred option among these. One special feature of the system is that it takes pictures of the classroom at the start and finish of every class to make sure everyone is present. After determining the number of students in each photograph, YOLO V3 separates the faces that are known and those that are unknown, creating distinct spreadsheets. Monthly email reminders are also sent to teachers, parents, and students. The system that has been put into place shows strong real-time performance in counting and detecting jobs, with excellent facial recognition accuracy and overall efficiency. Keywords:, OpenCV, Local Binary Pattern Histogram (LBPH), Real-time Tracking, Facial Analysis, You Only Look Once (YOLO V3), Firebase Database.