Speaker dependent visual word recognition by using sequential mouth shape codes | IEEE Conference Publication | IEEE Xplore