ASVspoof 2021: accelerating progress in spoofed and deepfake speech detection

Yamagishi, Junichi; Wang, Xin; Todisco, Massimiliano; Sahidullah, Md; Patino, José; Nautsch, Andreas; Liu, Xuechen; Lee, Kong Aik; Kinnunen, Tomi; Evans, Nicholas; Delgado, Héctor

doi:10.21437/asvspoof.2021-8

Cited by 149 publications

(92 citation statements)

References 0 publications

Supporting

Mentioning

Contrasting

Unclassified

Order By: Relevance

“…We followed the example in the Fairseq project toolkit [52] to extract feature representations from self-supervised wav2vec 2.0 pre-trained model. 2…”

Section: Pre-trainingmentioning

confidence: 99%

“…The differences between training, development and evaluation data can lead to substantial differences in detection performance. For the most recent ASVspoof 2021 Logical Access (LA) evaluation [2], the equal error rate (EER) of the best performing baseline solution increased from 0.55% for the development set to 9.26% for the evaluation set [2]. Submission results show better performance [3][4][5][6][7][8][9], but the fundamental gap between performance for development and evaluation data remains, indicating a persisting lack of generalisation.…”

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

Automatic speaker verification spoofing and deepfake detection using wav2vec 2.0 and data augmentation

Tak,

Todisco,

Wang

et al. 2022

Preprint

Self Cite

View full text Add to dashboard Cite

The performance of spoofing countermeasure systems depends fundamentally upon the use of sufficiently representative training data. With this usually being limited, current solutions typically lack generalisation to attacks encountered in the wild. Strategies to improve reliability in the face of uncontrolled, unpredictable attacks are hence needed. We report in this paper our efforts to use self-supervised learning in the form of a wav2vec 2.0 front-end with fine tuning. Despite initial base representations being learned using only bona fide data and no spoofed data, we obtain the lowest equal error rates reported in the literature for both the ASVspoof 2021 Logical Access and Deepfake databases. When combined with data augmentation, these results correspond to an improvement of almost 90% relative to our baseline system.

show abstract

“…We followed the example in the Fairseq project toolkit [52] to extract feature representations from self-supervised wav2vec 2.0 pre-trained model. 2…”

Section: Pre-trainingmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

Automatic speaker verification spoofing and deepfake detection using wav2vec 2.0 and data augmentation

Tak,

Todisco,

Wang

et al. 2022

Preprint

Self Cite

View full text Add to dashboard Cite

show abstract

“…ADD 2022 adopted six detection baseline systems. Motivated by the ASVspoof challenge [13], we use Gaussian mixture model (GMM), light convolutional neural network (LCNN) [18] and RawNet2 [19] to train baseline models. We modified the officially released source code 2 to build GMM, LCNN and RawNet2 classifiers.…”

Section: Detection Baselinesmentioning

confidence: 99%

“…The ASVspoof 2019 [12] consists of two tasks: LA and PA. There are three tasks in the ASVspoof 2021 [13]: *Corresponding author.…”

Section: Introductionmentioning

confidence: 99%

“…The ASVspoof challenges have played a key role in fostering spoofed speech detection research, which mainly aim to protect automatic speaker verification systems from manipulation. Although the audio deepfake detection task is included in the ASVspoof 2021 [13], it only involves compressed audio similar to the LA task. However, it ignores many challenging attacking situations in realistic scenarios.…”

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

ADD 2022: the First Audio Deep Synthesis Detection Challenge

Yi¹,

Fu²,

Tao³

et al. 2022

Preprint

View full text Add to dashboard Cite

Audio deepfake detection is an emerging topic, which was included in the ASVspoof 2021. However, the recent shared tasks have not covered many real-life and challenging scenarios. The first Audio Deep synthesis Detection challenge (ADD) was motivated to fill in the gap. The ADD 2022 includes three tracks: low-quality fake audio detection (LF), partially fake audio detection (PF) and audio fake game (FG). The LF track focuses on dealing with bona fide and fully fake utterances with various real-world noises etc. The PF track aims to distinguish the partially fake audio from the real. The FG track is a rivalry game, which includes two tasks: an audio generation task and an audio fake detection task. In this paper, we describe the datasets, evaluation metrics, and protocols. We also report major findings that reflect the recent advances in audio deepfake detection tasks.

show abstract

FMFCC-A: A Challenging Mandarin Dataset for Synthetic Speech Detection

Zhang

et al. 2022

Lecture Notes in Computer Science

View full text Add to dashboard Cite

ASVspoof 2021: accelerating progress in spoofed and deepfake speech detection

Cited by 149 publications

References 0 publications

Automatic speaker verification spoofing and deepfake detection using wav2vec 2.0 and data augmentation

Automatic speaker verification spoofing and deepfake detection using wav2vec 2.0 and data augmentation

ADD 2022: the First Audio Deep Synthesis Detection Challenge

FMFCC-A: A Challenging Mandarin Dataset for Synthetic Speech Detection

Contact Info

Product

Resources

About