Deteksi Deteksi Wajah Bermasker Berbasis Tensorflow-Keras untuk Pengendalian Gerbang Akses Masuk Menggunakan Raspberry Pi 4: Tensorflow-Keras Base Masked Face Detection For Access Control Gate Using Raspberry Pi 4 in Align With COVID-19 Health Protocol

Friendly, Friendly
Sembiring, Zakaria
Safitri, Habibi Ramdani

Jurnal Ilmu Komputer dan Sistem Komputer Terapan (JIKSTRA)
Vol. 2 No. 2 (2020): Edisi Oktober

Recent COVID-19 pandemic has cause change in human habit. Wearing a mask is a must health protocol according to Ministry of Health regulation. The regulation was announced since the pandemic shows significance rose in cases since May 2020. By using face recognition algorithm in Tensorflow by using CNN, this research purpose is to create an automatically access control gate that can filtered out people who wear mask and not wearing mask. The system will be implemented using Raspberry Pi 4 to control a servo that can be used to block a passageway/gate. Masked face detection in this research can detect face without mask up to 99.5% correct. While this program has a high capability in detection face without mask, this program didn’t have high capability in detecting face using mask which result only 93% correct and worsening when the face using mask having picture/motif on the surface of the mask to only 53% correct detection. Overall accuracy reach for 76%.