Distinguishability-Driven Voice Generation for Speaker Anonymization via Random Projection and GMM

Speaker anonymization effectively conceals speaker identity in speech signals to protect privacy. To address issues in existing anonymization systems, including reduced voice distinguishability, limited anonymized voices, reliance on an external speaker pool, and vulnerability to privacy leakage aga...

Ausführliche Beschreibung

Bibliografische Detailangaben

Hauptverfasser:	Chunxia Wang, Qiuyu Zhang, Yingjie Hu, Huiyi Wei
Format:	Artigo
Sprache:	Inglês
Veröffentlicht:	MDPI AG 2026-01-01
Schriftenreihe:	Big Data and Cognitive Computing
Schlagworte:	speaker anonymization privacy protection random projection GMM distinguishability
Online-Zugang:	https://www.mdpi.com/2504-2289/10/2/43
Tags:	Tag hinzufügen Keine Tags, Fügen Sie das erste Tag hinzu!