Attendance Management System Using Face Detection

This paper presents a Hybrid Multi-Stage Face Detection Algorithm that integrates traditional and deep learning methods for improved accuracy and efficiency. The process begins with Preprocessing and Enhancement to refine image quality. Fast Face Candidate Selection (Haar + HOG + SVM) quickly detects potential faces, followed by Precise Localization using MTCNN to refine detections and extract facial landmarks. Deep Learning Verification (RetinaFace/YOLO) eliminates false positives, ensuring reliability. Finally, Face Tracking (Kalman Filter + SORT) maintains consistency in video streams. This approach provides a robust and adaptable solution for real-world face detection applications.