Zobrazeno 1 - 10
of 10 032
pro vyhledávání: '"Todisco"'
Autor:
Polimeno, Laura, Todisco, Francesco, Mastria, Rosanna, De Giorgi, Milena, Fieramosca, Antonio, Pugliese, Marco, Ballarini, Dario, Grudinina, Anna, Voronova, Nina, Sanvitto, Daniele
Transition-metal dichalcogenides monolayers exhibit strong exciton resonances that enable intense light-matter interactions at room temperature (RT). However, the sensitivity of these materials to the surrounding environment and their intense interac
Externí odkaz:
http://arxiv.org/abs/2410.17693
We introduce 2D-Malafide, a novel and lightweight adversarial attack designed to deceive face deepfake detection systems. Building upon the concept of 1D convolutional perturbations explored in the speech domain, our method leverages 2D convolutional
Externí odkaz:
http://arxiv.org/abs/2408.14143
Autor:
Todisco, Massimiliano, Panariello, Michele, Wang, Xin, Delgado, Héctor, Lee, Kong Aik, Evans, Nicholas
We present Malacopula, a neural-based generalised Hammerstein model designed to introduce adversarial perturbations to spoofed speech utterances so that they better deceive automatic speaker verification (ASV) systems. Using non-linear processes to m
Externí odkaz:
http://arxiv.org/abs/2408.09300
Autor:
Wang, Xin, Delgado, Hector, Tak, Hemlata, Jung, Jee-weon, Shim, Hye-jin, Todisco, Massimiliano, Kukanov, Ivan, Liu, Xuechen, Sahidullah, Md, Kinnunen, Tomi, Evans, Nicholas, Lee, Kong Aik, Yamagishi, Junichi
ASVspoof 5 is the fifth edition in a series of challenges that promote the study of speech spoofing and deepfake attacks, and the design of detection solutions. Compared to previous challenges, the ASVspoof 5 database is built from crowdsourced data
Externí odkaz:
http://arxiv.org/abs/2408.08739
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 sanitised, other attributes such as the spoken content should be pres
Externí odkaz:
http://arxiv.org/abs/2408.04306
Autor:
Panariello, Michele, Tomashenko, Natalia, Wang, Xin, Miao, Xiaoxiao, Champion, Pierre, Nourtel, Hubert, Todisco, Massimiliano, Evans, Nicholas, Vincent, Emmanuel, Yamagishi, Junichi
The VoicePrivacy Challenge promotes the development of voice anonymisation solutions for speech technology. In this paper we present a systematic overview and analysis of the second edition held in 2022. We describe the voice anonymisation task and d
Externí odkaz:
http://arxiv.org/abs/2407.11516
Autor:
Todisco Francesco, Malureanu Radu, Wolff Christian, Gonçalves P. A. D., Roberts Alexander S., Mortensen N. Asger, Tserkezis Christos
Publikováno v:
Nanophotonics, Vol 9, Iss 4, Pp 803-814 (2020)
Light-matter interactions at the nanoscale constitute a fundamental ingredient for engineering applications in nanophotonics and quantum optics. In this regard, Mie resonances supported by high-refractive index dielectric nanoparticles have recently
Externí odkaz:
https://doaj.org/article/f5b7d72a9516431c98aae24bfbb1a304
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 equitable treatment. This expectation applies to biometric systems such as a
Externí odkaz:
http://arxiv.org/abs/2404.17810
Autor:
Tomashenko, Natalia, Miao, Xiaoxiao, Champion, Pierre, Meyer, Sarina, Wang, Xin, Vincent, Emmanuel, Panariello, Michele, Evans, Nicholas, Yamagishi, Junichi, Todisco, Massimiliano
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
Externí odkaz:
http://arxiv.org/abs/2404.02677
The vast majority of approaches to speaker anonymization involve the extraction of fundamental frequency estimates, linguistic features and a speaker embedding which is perturbed to obfuscate the speaker identity before an anonymized speech waveform
Externí odkaz:
http://arxiv.org/abs/2309.14129