Object Detection and Tracking using YOLOv8 and DeepSORT
Atul Kumar Yadav, Pratyush Kumar Chaturvedi, Shallu Rani
Abstract
Object detection and tracking are critical capabilities for computer vision systems. This research proposes a method for real-time and recorded video-based multiple object detection and tracking in videos, utilizing cutting-edge computer vision algorithms. YOLO, a high-performance convolutional neural network for object detection, and DeepSORT, an algorithm for separating object instances and matching detections across frames based on motion and appearance, are combined to create an object detection and tracking pipeline. The study’s findings show how well YOLO’s quick object recognition and DeepSORT’s dependable object tracking work together to provide accurate and instantaneous object monitoring. The suggested method has a lot of promise for use in areas like object detection, traffic control, and video surveillance, which will improve automation and situational awareness. The findings provided here pave the way for more investigation and practical use of these methods, providing opportunities for future advancements in the domains of artificial intelligence and computer vision. It could have a significant impact on a variety of businesses that rely on precise object perception and tracking.