One video is sufficient? Human activity recognition using active video composition.

Michael S. Ryoo, Wonpil Yu

Browse the full WACV paper archive.