21 papers
CLRM: Cross-Layer Relation Modeling for Audio Deepfake Detection
Yangyang Qu, Siqing Qin, Kong Aik Lee +1
Audio deepfake detection (ADD) aims to distinguish bonafide speech from synthetic or manipulated speech. Representations from different layers of self-supervised learning (SSL) spe…
Preemptive defense against re-identification attacks on voice anonymization via adversarial perturbation
Michele Panariello, Yibo Bai, Massimiliano Todisco +1
Voice anonymization (VA) is used to conceal the voice identities in speech data. Performance is usually estimated using automatic speaker verification (ASV) as a proxy to judge the…
Multi-sample Synthetic Supervision for Accent Conversion
Yangyang Qu, Michele Panariello, Massimiliano Todisco +1
Accent conversion (AC) requires changing accent while preserving speaker identity and linguistic content, yet parallel recordings are scarce. Speech synthesis provides an alternati…
Shared-State Local Translations for Training-Free Voice Conversion
Yangyang Qu, Michele Panariello, Massimiliano Todisco +1
In one-shot training-free voice conversion (VC), the source and reference utterances may contain different linguistic content, so reliable frame-level correspondence between them c…
Partial Accent-Control Editing in Frozen Speech Representations for Accent Conversion
Yangyang Qu, Michele Panariello, Massimiliano Todisco +1
Accent conversion is the task of modifying a speech recording so that it sounds closer to a target accent while preserving linguistic content and other speaker-related characterist…
PAID: Phoneme-Aligned Acoustic Units for Cross-Corpus L1-Background Classification
Yangyang Qu, Massimiliano Todisco, Michele Panariello +1
First-language (L1) background classification from second-language English speech aims to identify pronunciation-related acoustic cues that remain informative across speakers and c…
Why Do You Say It Like That? A Phoneme-level Framework for Explainable Speech Deepfake Detection
Anna Taylor, Michele Panariello, Massimiliano Todisco +3
As the accuracy of speech deepfake detection improves with the use of self-supervised representations such as wav2vec 2.0 and HuBERT, understanding why the speech is classified as…
Positive-Incentive Noise Predictor for Adversarial Purification in Speaker Verification
Yibo Bai, Sizhou Chen, Michele Panariello +5
Modern automatic speaker verification (ASV) systems are vulnerable to adversarial perturbations. Diffusion-based purification has recently shown strong effectiveness against such p…
Latent Secret Spin: Keyed Orthogonal Rotations for Blind Speech Watermarking in Anisotropic Latent Spaces
Emma Coletta, Massimiliano Todisco, Michele Panariello +2
We introduce Latent Secret Spin (LSS), a blind speech watermarking method based on geometric operations in codec latent space. Based upon orthogonal rotations to principal componen…
Evaluating voice anonymisation using similarity rank disclosure
Shilpa Chandra, Matteo Pettenò, Nicholas Evans +7
The evaluation of voice anonymisation remains challenging. Current practice relies on automatic speaker verification metrics such as the equal error rate (EER). Performance estimat…
Fair-Gate: Fairness-Aware Interpretable Risk Gating for Sex-Fair Voice Biometrics
Yangyang Qu, Massimiliano Todisco, Chiara Galdi +1
Voice biometric systems can exhibit sex-related performance gaps even when overall verification accuracy is strong. We attribute these gaps to two practical mechanisms: (i) demogra…
The Third VoicePrivacy Challenge: Preserving Emotional Expressiveness and Linguistic Content in Voice Anonymization
Natalia Tomashenko, Xiaoxiao Miao, Pierre Champion +7
We present results and analyses from the third VoicePrivacy Challenge held in 2024, which focuses on advancing voice anonymization technologies. The task was to develop a voice ano…
ASVspoof 5: Evaluation of Spoofing, Deepfake, and Adversarial Attack Detection Using Crowdsourced Speech
Xin Wang, Héctor Delgado, Nicholas Evans +8
ASVspoof 5 is the fifth edition in a series of challenges which promote the study of speech spoofing and deepfake detection solutions. A significant change from previous challenge…
MDD: a Mask Diffusion Detector to Protect Speaker Verification Systems from Adversarial Perturbations
Yibo Bai, Sizhou Chen, Michele Panariello +3
Speaker verification systems are increasingly deployed in security-sensitive applications but remain highly vulnerable to adversarial perturbations. In this work, we propose the Ma…
Reference-free Adversarial Sex Obfuscation in Speech
Yangyang Qu, Michele Panariello, Massimiliano Todisco +1
Sex conversion in speech involves privacy risks from data collection and often leaves residual sex-specific cues in outputs, even when target speaker references are unavailable. We…
The Risks and Detection of Overestimated Privacy Protection in Voice Anonymisation
Michele Panariello, Sarina Meyer, Pierre Champion +4
Voice anonymisation aims to conceal the voice identity of speakers in speech recordings. Privacy protection is usually estimated from the difficulty of using a speaker verification…
ASVspoof 5: Design, Collection and Validation of Resources for Spoofing, Deepfake, and Adversarial Attack Detection Using Crowdsourced Speech
Xin Wang, Héctor Delgado, Hemlata Tak +26
ASVspoof 5 is the fifth edition in a series of challenges which promote the study of speech spoofing and deepfake attacks as well as the design of detection solutions. We introduce…
2D-Malafide: Adversarial Attacks Against Face Deepfake Detection Systems
Chiara Galdi, Michele Panariello, Massimiliano Todisco +1
We introduce 2D-Malafide, a novel and lightweight adversarial attack designed to deceive face deepfake detection systems. Building upon the concept of 1D convolutional perturbation…
Malacopula: adversarial automatic speaker verification attacks using a neural-based generalised Hammerstein model
Massimiliano Todisco, Michele Panariello, Xin Wang +3
We present Malacopula, a neural-based generalised Hammerstein model designed to introduce adversarial perturbations to spoofed speech utterances so that they better deceive automat…
Preserving spoken content in voice anonymisation with character-level vocoder conditioning
Michele Panariello, Massimiliano Todisco, Nicholas Evans
Voice anonymisation can be used to help protect speaker privacy when speech data is shared with untrusted others. In most practical applications, while the voice identity should be…
A Comparison of Differential Performance Metrics for the Evaluation of Automatic Speaker Verification Fairness
Oubaida Chouchane, Christoph Busch, Chiara Galdi +2
When decisions are made and when personal data is treated by automated processes, there is an expectation of fairness -- that members of different demographic groups receive equita…