Multimodal speech recognition using mouth images from depth camera | IEEE Conference Publication | IEEE Xplore