* RESEARCH TOPICS [#h371a1d2]
- Security system using speaker recognition.
- Fast speaker adaptation technique for speaker-independent continuous speech recognition system.
- Hierarchical statistical model for fast speech recognition.
- Speech recognition system using neural network.
- Text-to-speech synthesis based on hidden Marcov model.
- Voice conversion using HMM-based speech synthesis technique.
- Very low bit rate (100 bit/s) speech coder.
- Individual classification using bimodal (audio-visual) speech towards robust security system.
- Text-to-audiovisual speech synthesis.
- Image recognition model for automatic lipreading.
- Automatic part extraction from facial image for individual classification.
