Graph-PIT: Generalized permutation invariant training for continuous separation of arbitrary numbers of speakers

Automatic transcription of meetings requires handling of overlapped speech, which calls for continuous speech separation (CSS) systems. The uPIT criterion was proposed for utterance-level separation with neural networks and introduces the constraint that the total number of speakers must not exceed...

Ausführliche Beschreibung

Bibliographische Detailangaben

Veröffentlicht in:	arXiv.org 2021-09
Hauptverfasser:	Thilo von Neumann, Kinoshita, Keisuke, Boeddeker, Christoph, Delcroix, Marc, Haeb-Umbach, Reinhold
Format:	Artikel
Sprache:	eng
Schlagworte:	Algorithms Channels Computer Science - Sound Criteria Graph coloring Neural networks Permutations Segments Separation Speech Stitching
Online-Zugang:	Volltext
Tags:	Tag hinzufügen Keine Tags, Fügen Sie den ersten Tag hinzu!