The College of Computer Science and Information Technology at University of Basrah organized a scientific lecture entitled "Real-Time Vehicle Detection, Tracking, and Behavioral Classification Using YOLO, DeepSORT, and LSTM." The lecture aimed to highlight the latest computer vision and deep learning technologies used in detecting, tracking, and classifying vehicle behavior in aerial videos by integrating detection and tracking models with specialized neural networks for analyzing temporal data.
The lecture, presented by researcher Zahraa Hussein Jumaa, included a review of the proposed methodology, which relies on the YOLOv8 model for vehicle detection and classification, the DeepSORT and ByteTrack algorithms for tracking, and the Attention-LSTM network for classifying vehicle behavior as stationary or moving based on motion characteristics extracted from tracking paths. The researcher also presented the experimental results demonstrating the efficiency of the proposed system, along with a comparison of the performance of tracking and classification algorithms using a range of standard metrics. The presentation concluded with a discussion of the key findings, challenges, and future prospects for developing intelligent traffic monitoring systems using drones.
