Dealing with Occlusion with Partial Visibility
|Author(s)||Ahmed Ridwanul Islam, Mohammad Saidur Rahman, Mohammad Saiedur Rahaman|
|Issue Date||June 01, 2011|
|Publishing Date||June 01, 2011|
|Keywords||Object tracking, Occlusion, Partial visibility|
Object tracking is a process that follows a specific object through some consecutive frames of images to determine the movement of that object relative to the other objects of those frames. Simply, tracking is the problem of estimating the trajectory of an object in the image plane as it moves around a scene. Depending on the tracking domain, a tracker can also provide partial and full object occlusions along with the other features. Here the problem with occlusion is being focused. It is experimented that if the color and texture features are applied then after occlusion object can be tracked. But sometimes it fails to detect the objects after occlusion. In this experiment we tried to include the partial visibility information of an object individually and along with these factors and has tried to proof that with this extra information we can project a better and more realistic trajectory of the object being tracked both before and after the occlusion. For this purpose a new and a very simple algorithm has been proposed which is able to track objects before and after occlusion even if the color and texture do not come up to scratch. Some experimental results are shown along with several case studies trough which the effectiveness of the proposed algorithm is explored.