Multi-view action recognition using bag-of-words and various fusion schemes