Automated Textual Descriptions for a Wide Range of Video Events with 48 Human Actions.

Patrick Hanckmann, Klamer Schutte, Gertjan J. Burghouts

Browse the full ECCV paper archive.