Szczegóły publikacji
Opis bibliograficzny
Real until proven fake – Source-Level Audio Deepfake Detection (with PIPNet) / Alicja MARTINEK // W: CIKM'25 : proceedings of the 34th ACM international conference on Information and Knowledge Management : November 10 - 14, 2025, Seoul, Republic of Korea / eds. Meeyoung Cha, [et al.]. — New York : Association for Computing Machinery, Inc. (ACM), cop. 2025. — ISBN: 979-8-4007-2040-6. — S. 6797–6800. — Bibliogr. s. 6800, Abstr. — Publikacja dostępna online od: 2025-11-10. — A. Martinek dod. afiliacja: - NASK - National Research Institute Warsaw, Poland
Autor
Słowa kluczowe
Dane bibliometryczne
| ID BaDAP | 165433 |
|---|---|
| Data dodania do BaDAP | 2026-01-15 |
| Tekst źródłowy | URL |
| DOI | 10.1145/3746252.3761659 |
| Rok publikacji | 2025 |
| Typ publikacji | materiały konferencyjne (aut.) |
| Otwarty dostęp |  |
| Creative Commons | |
| Wydawca | Association for Computing Machinery (ACM) |
| Konferencja | ACM International Conference on Information and Knowledge Management 2025 |
Abstract
The rapid development of synthetic speech technologies poses significant challenges to digital security and authenticity verification. This work investigates the use of prototype-based neural networks for detecting and classifying audio deepfakes by tracing them back to their generative source. Using time-frequency representations of speech (spectrograms, mel-spectrograms, MFCC), we evaluated model performance across multilingual and monolingual setups. Our results demonstrate that PIPNet reliably distinguishes between real and synthetic speech and effectively identifies the source TTS generator, making it a strong candidate for source-level attribution in audio deepfake detection.
