Pre-training of Speaker Embeddings for Low-latency Speaker Change Detection in Broadcast News.

Leda Sari, Samuel Thomas, Mark Hasegawa-Johnson, Michael Picheny

Browse the full ICASSP paper archive.