90
Voted
ICCV
2009
IEEE
16 years 1 months ago
2009
IEEE
In this paper, we present a Deformable Action Template
(DAT) model that is learnable from cluttered real-world
videos with weak supervisions. In our generative model,
an action ...