Supervised model training for overlapping sound events based on unsupervised source separation.

Toni Heittola, Annamaria Mesaros, Tuomas Virtanen, Moncef Gabbouj

Browse the full ICASSP paper archive.