A framework of object tracking based on STC algorithm

Accuracy and real-time are two important indicators of target tracking. While meeting the real-time requirements, the important factor that affects tracking accuracy is occlusion, especially severe occlusion and it is also a key index to evaluate the robustness of the tracking algorithm. In this paper, we proposed a framework of object tracking based on image sequences, which combines tracking algorithm with the anti-occlusion scheme consisting of loss detection, trajectory prediction and target recapture to make a solution of target tracking. The occlusion solution can judge the occlusion time accurately, save the useful target information for template matching which is an important way to deal with the occlusion, predict the trajectory while the target disappears and capture the target in time when it reappears. The experiments proved that the proposed framework can track rigid object effectively under the condition that the target is completely obscured for a while, which improves the accuracy of the algorithm while ensuring a real-time tracking speed.