Seo, S.
, Aulov, O.
, Godil, A.
and Mangold, K.
(2025),
EVALUATING IDENTITY LEAKAGE IN SPEAKER DE-IDENTIFICATION SYSTEMS, ICASSP 2026 - IEEE International Conference on Acoustics, Speech, and Signal Processing, Barcelona, ES, [online], https://tsapps.nist.gov/publication/get_pdf.cfm?pub_id=960507
(Accessed August 29, 2025)
Official websites use .gov
A .gov website belongs to an official government organization in the United States.
Secure .gov websites use HTTPS
A lock (
) or https:// means you’ve safely connected to the .gov website. Share sensitive information only on official, secure websites.
EVALUATING IDENTITY LEAKAGE IN SPEAKER DE-IDENTIFICATION SYSTEMS
Published
Author(s)
, , ,
Abstract
Speaker de-identification aims to conceal a speaker's identity while preserving intelligibility of the underlying speech. We introduce a benchmark that quantifies residual identity leak- age with three complementary error rates: equal error rate (EER), cumulative match characteristic (CMC) hit rate, and embedding-space similarity measured via canonical correla- tion analysis (CCA) and Procrustes analysis. Evaluation re- sults reveal that all state-of-the-art speaker de-identification systems leak identity information. The strongest system per- forms only slightly better than random guessing, while the weakest achieves a 45% hit rate within the top 50 candidates based on CMC. These findings highlight persistent privacy risks in current speaker de-identification technologies.Proceedings Title
ICASSP 2026 - IEEE International Conference on Acoustics, Speech, and Signal Processing
Conference Dates
May 4-8, 2026
Conference Location
Barcelona, ES
Pub Type
Conferences
Download Paper
Citation
Issues
If you have any questions about this publication or are having problems accessing it, please contact [email protected].
Created August 21, 2025, Updated August 29, 2025