← Search

Carlos Franzreb

2 accepted papers

2026

CONTENT LEAKAGE IN LIBRISPEECH AND ITS IMPACT ON THE PRIVACY EVALUATION OF SPEAKER ANONYMIZATION

ICASSP 2026poster

Speaker anonymization aims to conceal a speaker's identity, without considering the linguistic content. In this study, we reveal a weakness of Librispeech, the dataset that is commonly used to evaluate anonymizers: the books read by the Librispeech speakers are so distinct, that speakers can be iden…

Cited by 0SourcePDFScholar
2026

IMPROVING THE SPEAKER ANONYMIZATION EVALUATION’S ROBUSTNESS TO TARGET SPEAKERS WITH ADVERSARIAL LEARNING

ICASSP 2026poster

The current privacy evaluation for speaker anonymization often overestimates privacy when a same-gender target selection algorithm (TSA) is used, although this TSA leaks the speaker's gender and should hence be more vulnerable. We hypothesize that this occurs because the evaluation does not account…

Cited by 0SourcePDFScholar