Research Groups

In human interface Laboratory, a variety of researches on intelligent and human-oriented interface and statistical signal processing ar performed. The research area of intelligent and human-oriented interface covers 1) robust speech recognition for which large vocabulary continuous speech recognition systems, speaker recognition, speaker adaptation, utterance verification and robustness issues are studied, and 2) emotional speech synthesis in which acoustic modeling, natural sound synthesis and prosody control techniques are developed. On the other hand, in the area of statistical signal processing, we focus on signal enhancement in which noise suppression, echo cancellation, signal separation and blind deconvolution are included.

- Machine learning techniques for speech/audio signal processing
- Robust speech recognition in reverberant environment
- Emotional speech and singing voice synthesis
- Audio data hiding

- N. S. Kim, T. G. Kang, S. J. Kang, C. W. Han and D. H. Hong, "Speech feature mapping based on switching linear dynamic system," IEEE Transactions on audio, speech and language processing, Vol. 20, No. 2, pp. 620-631, Feb. 2012.
- N. S. Kim, J. S. Sung and D. H. Hong, "Factored MLLR adaptation," IEEE Signal Processing Letters, Vol. 18, No. 2, pp. 99-102, Feb. 2011.
- K. Cho, H. S. Yun and N. S. Kim, "Robust data hiding for MCLT based acoustic data transmission," IEEE Signal Processing Letters, Vol. 17, No. 7, pp. 679-682, Jul. 2010.