Real-Time Detection of Suspicious Behaviors and Theft Prevention Using an Artificial Intelligence Solution

Bedjboudja, Anas2025-11-252025-11-252025-06-30https://dspace.univ-tlemcen.dz/handle/112/25285This research develops a real-time shoplifting detection system using dual-stream deep learning that combines video analysis with human pose estimation. The system achieves 92.45% accuracy in detecting suspicious retail behaviors, enabling proactive theft prevention through automated surveillance and immediate alert generationenReal-Time Detection of Suspicious Behaviors and Theft Prevention Using an Artificial Intelligence SolutionReal-Time Detection of Suspicious Behaviors and Theft Prevention Using an Artificial Intelligence SolutionThesis