Improved action recognition by combining multiple 2D views in the bag-of-words model | IEEE Conference Publication | IEEE Xplore