Video Object Tracking
Maheshkumar H. Kolekar · 2018
Video object tracking is the process of estimating the positions of moving objects over time using a camera. A consistent label can be assigned by the tracker to the moving objects in different frames of a video. Further, for tracking objects, there is a need to recognize the object from the video frame. In video object tracking, occlusions significantly undermine the performance of any tracking algorithms. This chapter discusses the tracking approaches such as region-based tracking, contour-based tracking, feature-based tracking, and mean shift-based tracking. In region-based tracking, the deviations of the image sections are used for tracking moving objects. Contour-based tracking uses the boundary contour of a moving and deforming object to track it. To obtain the location and pose of the observer, model-based tracking is used if readers know the model of the environment. Mean shift is a non-parametric feature-space analysis technique that uses the given discrete data sample of a density function to locate its maxima.