This study develops a web-based digital attendance platform integrated with real-time facial recognition for PT Fiber Networks Indonesia. The system addresses the limitations of manual attendance logging, which is time-consuming during monthly recapitulation and vulnerable to data falsification. Developed using the iterative Spiral model—encompassing planning, risk analysis, engineering, and continuous evaluation phases—the platform's functional reliability was validated through Black Box Testing. Furthermore, quantitative evaluations across 120 biometric test trials from 12 registered employees demonstrated a high recognition accuracy rate of 95.8% with a 0% False Acceptance Rate (FAR). Under optimal lighting conditions, the average system latency from image capturing to database logging was achieved at 1.4 seconds per transaction. All verified attendance records are synchronously logged into a centralized database in real-time. Conclusively, this client-side web biometric architecture significantly eliminates fraudulent attendance practices and reduces administrative recapitulation time from hours to mere seconds, proving the technical viability of zero-installation biometric systems in dynamic industrial environments.
Copyrights © 2026