Efficient Facial Recognition: Leveraging VGGFace2 for Superior Accuracy and Performance

Abstract

Nowadays, facial recognition is becoming one of the most crucial technologies for attendance management, access control, and identity verification. Nevertheless, it becomes quite challenging to provide accurate recognition results in an efficient way. In this research work, the authors propose a face recognition framework that utilizes Multi-task Cascaded Convolutional Neural Networks (MTCNN) for face detection and alignment, VGGFace2 to extract deep facial features, Principal Component Analysis (PCA) to reduce the number of dimensions, Support Vector Machine (SVM) classifier, and ensemble classifier. The proposed framework was evaluated on Labeled Faces in the Wild Home (LFWH) dataset and compared with other recent face recognition methods. The experimental results show that the proposed method provides a recognition accuracy with Precision, Recall, F1-score and Area Under the Receiver Operating Characteristic Curve (ROC-AUC). Besides, the framework provides the False Acceptance Rate (FAR), False Rejection Rate (FRR), and Equal Error Rate (EER), which makes the method highly reliable and robust for face recognition.