Large scale continuous visual event recognition using max-margin Hough transformation framework | Publicación