CN-Celeb: Multi-genre speaker recognition

Li, Lantian; Liu, Ruiqi; Kang, Jiawen; Fan, Yue; Cui, Hao; Cai, Yunqi; Vipperla, Ravichander; Zheng, Thomas Fang; Wang, Dong

doi:10.1016/j.specom.2022.01.002

Cited by 59 publications

(25 citation statements)

References 35 publications

Supporting

Mentioning

Contrasting

Unclassified

Order By: Relevance

“…Experiments are carried out with the Voxceleb1+2 [16] and the CNCeleb1 databases [17]. A vanilla ResNet34 [18] model is trained with 1029K utterances from 5994 speakers in the training set of Voxceleb2.…”

Section: Methodsmentioning

confidence: 99%

Unifying Cosine and PLDA Back-ends for Speaker Verification

Peng¹,

He²,

Ding³

et al. 2022

Preprint

View full text Add to dashboard Cite

State-of-art speaker verification (SV) systems use a backend model to score the similarity of speaker embeddings extracted from a neural network model. The commonly used back-end models are the cosine scoring and the probabilistic linear discriminant analysis (PLDA) scoring. With the recently developed neural embeddings, the theoretically more appealing PLDA approach is found to have no advantage against or even be inferior the simple cosine scoring in terms of SV system performance. This paper presents an investigation on the relation between the two scoring approaches, aiming to explain the above counter-intuitive observation. It is shown that the cosine scoring is essentially a special case of PLDA scoring. In other words, by properly setting the parameters of PLDA, the two back-ends become equivalent. As a consequence, the cosine scoring not only inherits the basic assumptions for the PLDA but also introduces additional assumptions on the properties of input embeddings. Experiments show that the dimensional independence assumption required by the cosine scoring contributes most to the performance gap between the two methods under the domain-matched condition. When there is severe domain mismatch and the dimensional independence assumption does not hold, the PLDA would perform better than the cosine for domain adaptation.

show abstract

Section: Methodsmentioning

confidence: 99%

Unifying Cosine and PLDA Back-ends for Speaker Verification

Peng¹,

He²,

Ding³

et al. 2022

Preprint

View full text Add to dashboard Cite

show abstract

“…For all three tracks in the CNSRC 2022, we follow the same data usage setup. 797 speakers from CN-Celeb1 dev [1] and 1996 speaker from CN-Celeb2 [2] are used as the training data.…”

Section: Data Usagementioning

confidence: 99%

The SJTU X-LANCE Lab System for CNSRC 2022

Chen¹,

Liu²,

Han³

et al. 2022

Preprint

View full text Add to dashboard Cite

This technical report describes the SJTU X-LANCE Lab system for the three tracks in CNSRC 2022. In this challenge, we explored the speaker embedding modeling ability of deep ResNet (Deeper r-vector). All the systems are only trained on the Cnceleb training set and we use the same systems for the three tracks in CNSRC 2022. In this challenge, our system ranks the first place in the fixed track of speaker verification task. Our best single system and fusion system achieve 0.3164 and 0.2975 minDCF respectively. Besides, we submit the result of ResNet221 to the speaker retrieval track and achieve 0.4626 mAP.

show abstract

“…An x-vector embedding extractor [25] was pre-trained using CN-Celeb [26], a Mandarin speaker ID dataset. A Kaldi recipe was followed for training the x-vector model [27]. Inputs to the x-vector model were 30-dimensional MFCCs and the model consisted of 5 layers of time-dilated convolutional networks followed by average pooling and two fully connected layers; the size of the x-vector embedding was 512.…”

Section: X-vector Embedding With Cnn Classifiermentioning

confidence: 99%

FrAUG: A Frame Rate Based Data Augmentation Method for Depression Detection from Speech Signals

Ravi¹,

Wang²,

Flint³

et al. 2022

Preprint

View full text Add to dashboard Cite

In this paper, a data augmentation method is proposed for depression detection from speech signals. Samples for data augmentation were created by changing the frame-width and the frame-shift parameters during the feature extraction process. Unlike other data augmentation methods (such as VTLP, pitch perturbation, or speed perturbation), the proposed method does not explicitly change acoustic parameters but rather the time-frequency resolution of frame-level features. The proposed method was evaluated using two different datasets, models, and input acoustic features. For the DAIC-WOZ (English) dataset when using the DepAudioNet model and mel-Spectrograms as input, the proposed method resulted in an improvement of 5.97% (validation) and 25.13% (test) when compared to the baseline. The improvements for the CON-VERGE (Mandarin) dataset when using the x-vector embeddings with CNN as the backend and MFCCs as input features were 9.32% (validation) and 12.99% (test). Baseline systems do not incorporate any data augmentation. Further, the proposed method outperformed commonly used data-augmentation methods such as noise augmentation, VTLP, Speed, and Pitch Perturbation. All improvements were statistically significant.

show abstract

CN-Celeb: Multi-genre speaker recognition

Cited by 59 publications

References 35 publications

Unifying Cosine and PLDA Back-ends for Speaker Verification

Unifying Cosine and PLDA Back-ends for Speaker Verification

The SJTU X-LANCE Lab System for CNSRC 2022

FrAUG: A Frame Rate Based Data Augmentation Method for Depression Detection from Speech Signals

Contact Info

Product

Resources

About