Noise and room acoustics distorted speech recognition by HMM composition.

Satoshi Nakamura, Tetsuya Takiguchi, Kiyohiro Shikano

Browse the full ICASSP paper archive.