106
click to vote
CVPR
2011
IEEE
14 years 5 months ago
2011
IEEE
In this paper we explore the idea of using high-level semantic concepts, also called attributes, to represent human actions from videos and argue that attributes enable the constr...