Unsupervised segmentation of continuous speech using vector autoregressive time-frequency modeling errors.

Petri Korhonen, Unto K. Laine

Browse the full Interspeech paper archive.