Implementation of an Automated Identification System Using Facial Recognition Based on Principal Component Analysis (PCA)

Abstract

The administration of student attendance in higher institutions is a necessary responsibility that has conventionally relied on handwritten approaches, such as verbal roll calls and paper-based sign-in sheets. Although widely used, these traditional methods are time-consuming, minimize valuable instructional time, and are vulnerable to practices like proxy attendance, which compromise the accuracy and reliability of attendance records. To handle these limitations, this research implemented an intelligent student detection system based on facial recognition biometrics. The application was developed to generate a secure, contactless, and effective system necessary for identifying, capturing, and validating students' facial attributes in real time. The research adopted a quantitative experimental approach to design and assess the presented facial detection-based attendance model. The development embedded the Principal Component Analysis (PCA) techniques, generally known as the Eigenfaces approach, with the Python Flask web framework, OpenCV for computer vision tasks, and SQLite for data storage and administration. The performance results indicated the efficiency of the presented application. Testing the trained classifier on a stratified test dataset generated a macro accuracy of 95.8%, a precision of 96.3%, and an F1-score of 95.8%, demonstrating a strong level of classification performance. It is suggested that university management develop this biometric attendance platform across pilot deployments in controlled classroom ecosystems with standard lighting conditions.