PERANCANGAN DAN IMPLEMENTASI APLIKASI E-PRESENSI MAHASISWA BERBASIS GEOLOCATION DAN FACE RECOGNITION MENGGUNAKAN METODE PROTOTYPE

Abstract

Recording student classroom attendance represents a pivotal disciplinary metric governing final examination eligibility in higher education. Conventional manual attendance mechanisms suffer from widespread proxy signature fraud and classroom entry overhead. This research engineers an integrated mobile e-presence platform combining Haversine GPS Geolocation Geofencing with biometric Face Recognition built through the Prototype engineering paradigm. The biometric pipeline integrates the Haar Cascade Classifier for rapid face detection with Local Binary Pattern Histogram (LBPH) for texture feature extraction against training datasets. Built on Flutter and a Laravel RESTful API, functional black-box testing achieved a 100% success rate across 20 test cases. Biometric accuracy audits across 50 students under varying light conditions and head tilt poses attained a 94.80% matching accuracy with a 0.82-second inference latency. Geolocation validation demonstrated 100% efficacy in rejecting submissions outside a 50-meter radius while intercepting Mock GPS spoofing, compressing classroom attendance logging from 8 minutes to under 60 seconds.