Absensi RSUD Mangusada Bali — Biometric Face Recognition & Geofencing
High-reliability mobile attendance system featuring on-device facial recognition and anti-spoofing geofencing for RSUD Mangusada Bali.
Overview & Objectives
Absensi RSUD Mangusada Bali is an enterprise mobile attendance platform engineered to guarantee precise hospital staff presence through real-time on-device facial recognition and polygon geofencing coordinate verification around RSUD Mangusada Bali.
[!] Problem Statement
Conventional hospital attendance methods caused physical bottleneck lines during medical shift handovers, while standard mobile attendance apps remained vulnerable to mock GPS spoofing and camera photo replay exploits.
[✓] Engineering Solution
Engineered a streamlined Flutter application with real-time on-device biometric vector inference, verified against hospital polygon geofencing boundaries and cryptographically signed payload timestamps.
System Architecture
Key Engineering Challenges & Solutions
Challenge 1: Inference latency on entry-level Android hardware caused camera preview stuttering.
Resolution: Decoupled image frame acquisition from inference pipeline by running vector embeddings inside dedicated Dart background isolates with frame throttling.
Challenge 2: Fake GPS tools circumventing native OS location reporting.
Resolution: Implemented multi-layer root detection, mock location provider verification, and cell tower triangulation cross-checking.
Key Results & Verification
- 99.4% biometric verification accuracy under standard lighting conditions
- Under 450ms total verification and checkout latency
- Zero reported proxy attendances across hospital operational deployment