Multi-Camera Person Tracking: Integrating YOLOv8 with ByteTrack
Nirali Anand Pandya, Narendrasinh C. Chauhan · International Journal of Electrical and Electronics Engineering · 2024
Accurate and efficient person tracking in complex, multi-camera environments remains challenging. This paper proposes a novel approach that integrates the strengths of YOLOv8, an advanced model for object detection, with ByteTrack, an advanced multi-object tracking algorithm. The proposed framework is evaluated on the challenging Multi-camera Pedestrians Video Dataset to assess its performance in complex real-world scenarios. Experimental results demonstrate the effectiveness of the proposed method in accurately tracking pedestrians across multiple cameras, outperforming existing state-of-the-art techniques. Integrating YOLOv8 and ByteTrack enables robust pedestrian detection and tracking, even in challenging conditions such as occlusions, varying illumination, and camera perspectives. The proposed approach holds significant potential for intelligent surveillance systems, crowd analysis, and autonomous vehicle applications.