Facial and Geofencing-Based Attendance Tracking System for Deployed Personnel

Abstract

This study explored developing and deploying a facial and geofencing-based
attendance tracking system designed for the Cagayan de Oro City Police Office
(COCPO), Philippines. Its main goal was to improve the accuracy of patrol tracking
and officer accountability. The system, built with Python, combined facial recognition
with GPS/Location-Based Services (LBS) and the Haversine formula for precise
geofence validation to automate attendance tracking through a mobile app and web
dashboard. Officers authenticated themselves by facial recognition on smartphones
and had to take a real-time selfie only after confirming their presence within a
designated patrol zone via geofence validation. Usability testing with COCPO officers
showed a facial recognition accuracy of 99% under optimal conditions, a geofencing
accuracy of 92.18% in open areas (dropping to 71.33% in urban environments), and
a System Usability Scale (SUS) score of 85.56. The results demonstrated that the
system greatly enhanced accountability and transparency, resolving issues associated
with manual record-keeping. This research offers a validated dual-authentication
framework for law enforcement, showcasing its potential scalability for public safety
uses.