Track to the future: Spatio-temporal video segmentation with long-range motion cues.

Jos Lezama, Karteek Alahari, Josef Sivic, Ivan Laptev

Browse the full CVPR paper archive.