Text this: Video Scene Detection Using Compact Bag of Visual Word Models