All-phoneme ergodic hidden Markov network for unsupervised speaker adaptation.

Yasunaga Miyazawa, Jun-ichi Takami, Shigeki Sagayama, Shoichi Matsunaga

Browse the full ICASSP paper archive.