Graph Based ANPR in Presence of Occlusion

Vehicle theft remains a significant global concern, resulting in substantial financial losses and safety risks. The authors propose Vision Shield, a robust system utilizing deep learning methodologies to improve vehicle tracking and theft detection through automated surveillance. The system integrates computer vision techniques and real-time analytics to monitor and track vehicles across diverse locations effectively. Experimental results highlight the system’s ability to perform precise vehicle identification, maintain low false positive rates, and achieve high recall, thereby supporting timely interventions. The findings underscore the adaptability of Vision Shield to various environments, demonstrating its scalability for broader applications.