Multiple Vehicle Detection and Tracking via RetinaNet with KLT Tracker

This study presents a novel approach to vehicle detection and tracking using the RetinaNet architecture integrated with the KLT tracker. The proposed method achieves a detection accuracy of 93% and tracking accuracy of 89%, demonstrating significant improvements over traditional methods. Our approach is validated through extensive experiments, highlighting its robustness and applicability in real-world scenarios. The contributions of this research lie in the innovative combination of deep learning and tracking algorithms, offering a scalable and effective solution for real-time vehicle monitoring. This hybrid framework addresses challenges in occlusion, varying lighting conditions, and multi-object detection, making it a promising advancement in the field of intelligent transportation systems.