Rancang Bangun Sistem Absensi Otomatis Berbasis Pengenalan Wajah Menggunakan Model CNN Pretrained pada Platform Web

Armando, Gali
Simangunsong, Marta Aulia
Mediansyah, Teguh Arif
Harahap, Zulkaidah
Rahelta, Cristina Elseria
Hutahean, Harvei Desmon
Syahputra, Fahmy
Sabrina, Elsa

QISTINA: Jurnal Multidisiplin Indonesia
Vol 4, No 2 (2025): December 2025

Conventional attendance methods often lead to queues, time inefficiency, and potential violation of health protocols, necessitating a fast, non-contact, and real-time attendance recording system. This research aims to design and implement a web-based attendance system as a local prototype using face recognition biometrics. The system was developed using Python with the Flask Framework and OpenCV. The core face recognition process combines Dlib's Pretrained CNN model for 128-dimensional feature vector extraction (face embedding) and the K-NN method for classification based on Euclidean Distance calculation. Testing results indicate that the system successfully performs accurate and real-time facial identification. The system is capable of automatically logging attendance times, providing audio feedback, and storing the attendance data recapitulation in an Excel (.xlsx) file. Thus, this system provides an effective and efficient non-contact attendance solution.