Design and Development of Face Recognition System for Boarding House using LBPH Algorithm and Firebase (Case Study: Kos Putra Muslim)

The challenges of monitoring unregistered guests in boarding houses often pose security risks and financial losses for the owners. This research answers this challenge by designing and building a fully integrated automated facial recognition system. This system is implemented on the Raspberry Pi 5 using the Local Binary Pattern Histogram (LBPH) algorithm due to its efficiency. The architecture of this system is supported by a Flask-based REST API that allows dynamic management of facial data and device control through an Android application. In contrast to previous research, the novelty of this system lies in the interactive data management mechanism and centralized control of devices through applications, which have not existed in similar studies. The performance of the system was evaluated using quantitative experimental methods to measure the success rate on various test scenarios and compared with the Dlib algorithm. The results showed that all system functionality was running as designed, where the LBPH algorithm proved to be reliable under ideal conditions but significantly less robust than Dlib under more dynamic conditions. It was concluded that this prototype succeeded in becoming a more transparent functional solution for the security of rental housing.