This research aims to design and implement a home security system based on facial
recognition that is capable of working in real-time and integrated with the Internet of Things
(IoT), in order to overcome the limitations of conventional security systems that are not yet able
to detect and respond to potential threats automatically and still rely on manual supervision.
The system was developed using an ESP32-CAM module with a Local Binary Pattern
Histogram (LBPH) algorithm for the facial identification process, and is integrated with a
buzzer and the Telegram application as a two-layer warning system. The method used is
prototyping with an iterative approach for two months through direct testing in a residential
environment. The test results show that the system is able to recognize faces with 95.42%
accuracy, provides a fast response, and works stably in various lighting conditions. The
conclusion of this research shows that the system is effective, economical, and can be
implemented without major changes to the building structure. This system also shows potential
for further development on a more complex smart home scale.