Application of Arcface and support vector machine in face recognition

Tóm tắt

Face recognition is one of the key domains of artificial intelligence (AI) and computer vision, widely applied in security, surveillance, and modern smart devices. In Vietnam, most existing systems are costly to deploy and have limited customization capabilities for practical needs. In this study, we propose a face recognition model that combines ArcFace for deep feature extraction and Support Vector Classification (SVC) for effective classification. The system is fully developed on the open- source Python platform, integrating OpenCV, NumPy, Scikit-learn, and a MySQL database for systematic user information management. Experiments conducted on 1,200 images from three user classes under diverse environmental conditions demonstrate that the model achieves an accuracy of 73.3%. In addition, the system successfully implements real-time recognition with detailed user information display. The results confirm the feasibility of building an efficient and low-cost face recognition system, well-suited to application conditions in Vietnam.