|
Review A SYSTEMATIC LITERATURE REVIEW OF FACE RECOGNITION BASED ATTENDANCE SYSTEM: METHODS, TRENDS, AND RESEARCH GAPSPages 135-154
Abstract
This systematic literature review examines facial-recognition-based attendance systems published between 2020 and 2024. The review analyzes preprocessing techniques, face detection approaches, feature extraction methods, recognition strategies, evaluation metrics, and state-of-the-art implementations across educational and workplace environments. The findings show that traditional approaches such as Viola–Jones, Haar Cascade, LBP, and LBPH remain widely used due to their computational efficiency, while deep learning models, including CNN-based architectures, FaceNet, and MobileFaceNet, offer superior accuracy but require higher computational resources. Single-face recognition systems demonstrate high accuracy yet lack scalability for real classroom scenarios, whereas multi-face recognition systems improve efficiency but still rely heavily on conventional methods with limited robustness under pose and illumination variations. The study highlights ongoing challenges in scalability, environmental resilience, and evaluation consistency, providing insights for enhancing future facial-recognition-based attendance solutions.
Keywords: Facial recognition, Attendance system, Multi-face detection, Feature extraction, Systematic literature review.
|
Agustiyar Agustiyar, 
PDF 