Improving Rapid Unsupervised Speaker Adaptation Based On Hmm Sufficient Statistics.

Randy Gomez, Tomoki Toda, Hiroshi Saruwatari, Kiyohiro Shikano

Browse the full ICASSP paper archive.