A new phonetic tied-mixture model for efficient decoding.

Akinobu Lee, Tatsuya Kawahara, Kazuya Takeda, Kiyohiro Shikano

Browse the full ICASSP paper archive.