Audio-visual modeling for bimodal speech recognition | IEEE Conference Publication | IEEE Xplore