Tracking an Object in Video using the Mean Shift Protocol

Abstract

In computer vision and remote sensing, object tracking is an emotive subject. If only the target's starting state (in a frame of video) exists, object tracking usually employs a bounding box that holds onto the area of interest (AOI). In this method, a rectangular target window appears in a starting frame for a target that is moving in a video, and then the tracked object is distinguished from the backdrop by analysing the data inside that window. It happens to be the main goal of object tracking to determine the exact location of the object in visuals in an uninterrupted way and accurately against scenes that are changing.