Model adaptation based on HMM decomposition for reverberant speech recognition.

Tetsuya Takiguchi, Satoshi Nakamura, Qiang Hou, Kiyohiro Shikano

Browse the full ICASSP paper archive.