Deep Neural Network-based Single Object Tracking

In this paper, we put forward the notion of an approach centered on single object tracking. The single object tracker is going to find one object, and then it is going to track that object over the whole frame of the video. The basic elements of this methodology are images, groundtruths, neural network, and detector which are used to make a single object tracker. The neural network used for this tracking method is RESNET-101. Other trackers are also efficient in tracking the object, but still not getting accurate predicted bounding boxes on the selected object, this field gives other people a chance to make different trackers that can do perfect tracking. The datasets used in this paper are the Online object tracking benchmark(OOTB) and Unmanned Aerial Vehicle(UAV).