Innovative and Efficient Face Recognition System Utilizing VGGFace-16 with Multiple Classifiers
R. Janarthanan, Jay Gandhi, V. Prabakaran, Parimala Veluvali · 2025
Face recognition plays a vital role in modern applications such as surveillance, security systems, and user authentication. Traditional approaches often struggle with variations in pose, lighting, and occlusion, limiting their effectiveness in real-world scenarios. To address these challenges, this research proposes a robust and efficient face recognition framework that combines the deep feature extraction capabilities of the VGGFace-16 architecture with an ensemble of three classical classifiers: Support Vector Machines (SVM), Random Forest, and K-Nearest Neighbors (KNN). Features extracted from VGGFace-16 are classified using the three methods, and the final decision is made through a majority voting mechanism to ensure accuracy and stability. The proposed system is evaluated on two benchmark datasets—Labeled Faces in the Wild (LFW) and CASIA-WebFace—demonstrating superior performance in terms of accuracy, precision, recall, and F1-score when compared to individual classifier models. The experimental results validate the effectiveness and computational efficiency of the proposed framework. This study concludes that the integration of deep learning with an ensemble of traditional classifiers enhances both recognition accuracy and adaptability, making the system well-suited for deployment in resource-constrained and rea-ltime environments.