Tube ConvNets: Better exploiting motion for action recognition.

Zhihao Li, Wenmin Wang, Nannan Li, Jinzhuo Wang

Browse the full ICIP paper archive.