Text-to-visual speech synthesis based on parameter generation from HMM | IEEE Conference Publication | IEEE Xplore