Abstract: Consistency in 3D tracking is critical for vision-based autonomous driving systems. However, most existing multi-object tracking (MOT) methods are confronted with the limitation: detectors ...

Abstract: 6D object tracking plays an important role in various applications, including robotic manipulation and virtual reality. While current methodologies have achieved significant advancements ...

A real-time computer vision system that monitors objects using YOLOv8 and automatically detects when registered objects go missing from the surveillance area. Built with Python, OpenCV, and a modern ...

Real-time small object detection and tracking on edge devices using EdgeYOLO-S, Norfair, ByteTrack and TensorRT. This repository contains a fine-tuned EdgeYOLO-S model for small object tracking using ...