SISTEM PENDETEKSI WAJAH DENGAN METODE VIOLA JONES MENGGUNAKAN ESP32-CAM

Abstract

Increased security in rooms that are confidential or require high privacy, is very much
needed. One way to increase this security is to use a security system that utilizes the biological
characteristics of facial images that are owned by every human being. The biological
characteristics of human facial images can be a solution for security systems in the room
because human facial images are difficult to imitate and modify by others. This study aims to
implement a security system for detecting human face images using the Viola Jones method
with the Haar Classification Feature. The Viola Jones method with the Haar Classification
feature is suitable as a good visual recognition when detecting facial images via the ESP32-
Cam microcontroller and for streaming video using a OV2640 camera. The test results in this
study showed that of the 4 objects used in human face images, the ideal distance when
detecting facial images was 30 cm, with a 100% success rate. While the level of slope tolerance
that has been detected is a maximum with a slope of 30° to the position of the face facing
straight ahead.