Online multi-modal task-driven dictionary learning and robust joint sparse representation for visual tracking | IEEE Conference Publication | IEEE Xplore