Learning motion categories using both semantic and structural information (2007)

by S Wong, T Kim, R Cipolla
Venue:Proc. IEEE International Conference on Computer Vision and Pattern Recognition