SEF-MK: Speaker-Embedding-Free Voice Anonymization through Multi-k-means Quantization
Fuente:
arXiv
Saved in:
| Main Authors: | , , , |
|---|---|
| Format: | Preprint |
| Published: |
2025
|
| Subjects: | |
| Online Access: | |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1866911106705915904 |
|---|---|
| author | Tang, Beilong Miao, Xiaoxiao Wang, Xin Li, Ming |
| author_facet | Tang, Beilong Miao, Xiaoxiao Wang, Xin Li, Ming |
| contents | Voice anonymization protects speaker privacy by concealing identity while preserving linguistic and paralinguistic content. Self-supervised learning (SSL) representations encode linguistic features but preserve speaker traits. We propose a novel speaker-embedding-free framework called SEF-MK. Instead of using a single k-means model trained on the entire dataset, SEF-MK anonymizes SSL representations for each utterance by randomly selecting one of multiple k-means models, each trained on a different subset of speakers. We explore this approach from both attacker and user perspectives. Extensive experiments show that, compared to a single k-means model, SEF-MK with multiple k-means models better preserves linguistic and emotional content from the user's viewpoint. However, from the attacker's perspective, utilizing multiple k-means models boosts the effectiveness of privacy attacks. These insights can aid users in designing voice anonymization systems to mitigate attacker threats. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2508_07086 |
| institution | arXiv |
| publishDate | 2025 |
| record_format | arxiv |
| spellingShingle | SEF-MK: Speaker-Embedding-Free Voice Anonymization through Multi-k-means Quantization Tang, Beilong Miao, Xiaoxiao Wang, Xin Li, Ming Sound Machine Learning Audio and Speech Processing Voice anonymization protects speaker privacy by concealing identity while preserving linguistic and paralinguistic content. Self-supervised learning (SSL) representations encode linguistic features but preserve speaker traits. We propose a novel speaker-embedding-free framework called SEF-MK. Instead of using a single k-means model trained on the entire dataset, SEF-MK anonymizes SSL representations for each utterance by randomly selecting one of multiple k-means models, each trained on a different subset of speakers. We explore this approach from both attacker and user perspectives. Extensive experiments show that, compared to a single k-means model, SEF-MK with multiple k-means models better preserves linguistic and emotional content from the user's viewpoint. However, from the attacker's perspective, utilizing multiple k-means models boosts the effectiveness of privacy attacks. These insights can aid users in designing voice anonymization systems to mitigate attacker threats. |
| title | SEF-MK: Speaker-Embedding-Free Voice Anonymization through Multi-k-means Quantization |
| topic | Sound Machine Learning Audio and Speech Processing |
| url | https://arxiv.org/abs/2508.07086 |