Correlation tracking via robust region proposals

Recently, correlation filter-based trackers have received extensive researching interest because of their simplicity and superior running speed. However, such trackers perform poorly when the target undergoes occlusion, viewpoint change, or other challenging attributes due to pre-defined sampling strategy. To tackle these issues, the authors propose an adaptive region proposal strategy to facilitate object tracking. To be more specific, a novel tracking monitoring indicator is advocated to forecast tracking failure. Afterwards, detection and scale proposals were incorporated, respectively, to recover from model drift as well as handle aspect ratio variation. The algorithm was tested on several challenging video sequences, which demonstrates that the proposed algorithm outperforms other state-of-the-art trackers.