Pith. sign in

REVIEW 12 cited by

The VoicePrivacy 2024 Challenge Evaluation Plan

Not yet reviewed by Pith; the record is open.

Re-run · record.json Download PDF Read on arXiv ↗

This paper has not been read by Pith yet. Machine review is queued; the pith claim, tier, and objections will appear here once it completes.

SPECIMEN: schema-true, not a live event

T0 review · schema-true

One-sentence machine reading of the paper's core claim.

pith:XXXXXXXX · record.json · timestamp

arxiv 2404.02677 v2 pith:7WHVPB4Y submitted 2024-04-03 eess.AS cs.CLcs.CR

Natalia Tomashenko , Xiaoxiao Miao , Pierre Champion , Sarina Meyer , Xin Wang , Emmanuel Vincent , Michele Panariello , Nicholas Evans
show 2 more authors
Junichi Yamagishi Massimiliano Todisco
This is my paper
classification eess.AScs.CLcs.CR
keywords evaluationanonymizationchallengeparticipantssystemsdataorganizersresults
verification ladder T0 review T1 audit T2 compute T3 formal
0 comments
read the original abstract

The task of the challenge is to develop a voice anonymization system for speech data which conceals the speaker's voice identity while protecting linguistic content and emotional states. The organizers provide development and evaluation datasets and evaluation scripts, as well as baseline anonymization systems and a list of training resources formed on the basis of the participants' requests. Participants apply their developed anonymization systems, run evaluation scripts and submit evaluation results and anonymized speech data to the organizers. Results will be presented at a workshop held in conjunction with Interspeech 2024 to which all participants are invited to present their challenge systems and to submit additional workshop papers.

Discussion (0). Sign in to comment.

Forward citations

Cited by 12 Pith papers

Reviewed papers in the Pith corpus that reference this work. Sorted by Pith novelty score. Full citation record

  1. The Risks and Detection of Overestimated Privacy Protection in Voice Anonymisation

    eess.AS 2025-07 conditional novelty 7.0 of 10

    Mismatches between the anonymisation systems used to train and test speaker verification attacks can overestimate privacy protection, and a validation-vs-test gap can detect such mismatches.

  2. Content Anonymization for Privacy in Long-form Audio

    cs.SD 2025-10 conditional novelty 6.0 of 10

    LLM-based rewriting of transcripts inside an ASR-TTS pipeline defeats content-based speaker re-identification in long-form audio, driving attacker accuracy to chance.

  3. VoxGuard: Evaluating User and Attribute Privacy in Speech via Membership Inference Attacks

    cs.CR 2025-09 conditional novelty 6.0 of 10

    Evaluating voice anonymization at low false-positive rates reveals much stronger membership inference and attribute leakage than Equal Error Rate reports.

  4. SegReConcat: A Data Augmentation Method for Voice Anonymization Attack

    cs.SD 2025-08 conditional novelty 6.0 of 10

    SegReConcat, a word-shuffle-and-concatenate augmentation, improves attacker speaker verification against five of seven voice anonymization systems in the VPAC 2024 benchmark.

  5. Use Cases for Voice Anonymization

    eess.AS 2025-08 unverdicted novelty 6.0 of 10

    Voice anonymization should be designed and evaluated per use case; this paper proposes the first taxonomy of use cases plus requirements derived from a literature review and a public user study.

  6. Exploiting Context-dependent Duration Features for Voice Anonymization Attack Systems

    cs.SD 2025-07 conditional novelty 6.0 of 10

    A context-dependent encoding of phoneme durations identifies speakers far better than average-duration vectors and remains effective on anonymized speech without retraining on anonymized data.

  7. First Steps Towards Voice Anonymization for Code-Switching Speech

    eess.AS 2025-07 conditional novelty 6.0 of 10

    A new benchmark shows that a multilingual anonymization system preserves utility on Mandarin-English and Spanish-English code-switching speech, while language-independent baselines fail to protect privacy.

  8. Multimodal Speaker Verification as a Threat to Speaker Anonymization

    eess.AS 2026-07 conditional novelty 5.0 of 10

    Multi-utterance, audio-plus-text speaker verification lowers equal error rates on anonymized speech, leaving residual speaker-identifying information after voice anonymization.

  9. SEF-MK: Speaker-Embedding-Free Voice Anonymization through Multi-k-means Quantization

    cs.SD 2025-08 conditional novelty 5.0 of 10

    Randomly switching among multiple k-means quantizers on SSL speech features improves utility but weakens privacy against attackers who also use multiple quantizers.

  10. SecureSpeech: Prompt-based Speaker and Content Protection

    cs.SD 2025-07 conditional novelty 5.0 of 10

    A dual anonymization pipeline that uses an LLM to replace sensitive entities and a prompt-driven TTS to generate speech with a new, unrelated voice.

  11. Mitigating Language Mismatch in SSL-Based Speaker Anonymization

    eess.AS 2025-07 conditional novelty 5.0 of 10

    Fine-tuning an SSL content encoder on Japanese, especially when the encoder is pre-trained multilingually, makes anonymized Japanese and Mandarin speech much more intelligible while keeping speaker privacy at usable levels.

  12. Speaker Privacy and Security in the Big Data Era: Protection and Defense against Deepfake

    eess.AS 2025-09 accept novelty 1.0 of 10

    A concise survey of voice anonymization, deepfake detection, and speech watermarking as defenses against deepfake speech, with current challenges.

Pith tools