Audio-visual speech recognition using deep bottleneck features and high-performance lipreading | IEEE Conference Publication | IEEE Xplore