Learning hierarchical invariant spatio-temporal features for action recognition with independent subspace analysis

   Abstract