Human Face and Action Recognition Through CCTV Surveillance

Abstract: This project focuses on developing a system for human face and action recognition through CCTV surveillance, leveraging deep learning algorithms. By uploading CCTV footage videos and individual photos of persons of interest, the system aims to detect, track, and recognize faces and actions in real-time. The output provides the identified person's face, recognized actions, and a unique identifier along with timestamps indicating when the action occurred. Key components of the system include the YOLO v8 algorithm for object detection, Deep SORT algorithm for object tracking, and FaceSDK for face detection and recognition. The integration of these advanced technologies aims to provide a comprehensive solution for enhancing security measures and facilitating forensic analysis in surveillance environments. Through the utilization of deep learning techniques, the project contributes to advancing the capabilities of CCTV surveillance systems in recognizing and analysing human activities effectively.

Keywords: Human face recognition, Action recognition, CCTV surveillance, Deep learning, YOLO v8, Object detection, Deep SORT, Object tracking, FaceSDK, Forensic analysis, Security measures, Timestamp, Facial detection.