Object Detection and Tracking for any Video Using R-CNN

Abstract

When it comes to multiple object tracking, this research investigates a pragmatic method in which the primary emphasis is on associating objects effectively for online and real-time applications. Accordingly, detection quality is recognized as a critical component that influences tracking performance, with modifying the detector resulting in an improvement of up to 18.9 percent in tracking. Despite the fact that this approach just applies a basic combination of well-known methods such as the Kalman Filter as well as This time, we're using the Hungarian algorithm to monitor components, and the tracking accuracy we're getting is on par with the best online trackers. The tracker refreshes at a rate of more than 20 times quicker than many current state-of-the-art trackers due to the simplicity of the technology utilised in its design.