VoxCeleb: A Large-Scale Speaker Identification Dataset
Open accessThe DOI matches its Crossref record; 10 of the 10 references it cites were found in a registry or an index; none of the 16 sources its claims cite is checked for support yet.
Abstract
Most existing datasets for speaker identification contain samples obtained under quite constrained conditions, and are usually hand-annotated, hence limited in size. The goal of this paper is to generate a large scale text-independent speaker identification dataset collected 'in the wild'. We make two contributions. First, we propose a fully automated pipeline based on computer vision techniques to create the dataset from open-source media. Our pipeline involves obtaining videos from YouTube; performing active speaker verification using a two-stream synchronization Convolutional Neural Network (CNN), and confirming the identity of the speaker using CNN based facial recognition. We use this pipeline to curate VoxCeleb which contains hundreds of thousands of 'real world' utterances for over 1,000 celebrities. Our second contribution is to apply and compare various state of the art speaker identification techniques on our dataset to establish baseline performance. We show that a CNN based architecture obtains the best performance for both identification and verification.
Overview
CoChat Summary
TL;DR
The paper introduces VoxCeleb, a fully automated, large-scale speaker identification dataset built from YouTube videos “in the wild,” with 1,251 celebrities and over 100,000 utterances. It also benchmarks speaker recognition methods and reports that a CNN performs best on both identification and verification.
Key findings
- VoxCeleb contains over 100,000 utterances for 1,251 celebrities; the dataset is gender balanced, with 55% male speakers.
- The automated collection pipeline uses active speaker verification and face verification and was tuned to be highly precise: precision = 1.000 for both stages, with recall = 0.613 for active speaker verification and 0.726 for face verification.
- For speaker identification, the CNN achieves 80.5% top-1 and 92.1% top-5 accuracy, compared with 49.0% / 56.6% for I-vectors + SVM and 60.8% / 75.6% for I-vectors + PLDA + SVM.
- For speaker verification, the best reported result is CNN + Embedding with C min det = 0.71 and EER = 7.8%, better than I-vectors + PLDA (0.73, 8.8%) and GMM-UBM (0.80, 15.0%).
Methods
- Built VoxCeleb from YouTube videos using a fully automated multi-stage pipeline: candidate POI list, video download, face tracking, active speaker verification via SyncNet, and face verification via VGG Face CNN.
- Dataset scale/statistics reported: 1,251 POIs, 690 male POIs, 36 / 18 / 8 videos per POI (max/avg/min), 250 / 123 / 45 utterances per POI (max/avg/min), and 145.0 / 8.2 / 4.0 s utterance length (max/avg/min).
- Trained and evaluated a CNN on spectrograms with minimal preprocessing; identification used a 1,251-way classification setup, while verification compared I-vectors/PLDA, GMM-UBM, and CNN embedding approaches.
Limitations
- The paper states that the pipeline is designed to be conservative, and that this means “a lot of the downloaded videos” are discarded to reduce label errors.
- The paper notes that the collection starts from a list of 2,622 identities from VGG Face and uses YouTube search with “interview” appended, which is an evident source of selection bias toward publicly available, interview-like media.
- The dataset is described as coming from open-source media / YouTube, so it is limited to speakers and contexts available there; this is evident from the text, though not stated as a formal limitation.
Model-written summary. Check it against the paper before relying on it.
Integrity signals
Registry
Cites
Statistics
Verification
Claims the paper makes
0 Valid · 1 Not valid · 12 Not checked · 37 Own results and interpretations
In the paper
“| Face verification | [30] | 1.000 | 0.726 |”
What the sources say
- Not checked yetcites [30] [30] O. M. Parkhi, A. Vedaldi, and A. Zisserman, 'Deep face recognition,' in Proceedings of the British Machine Vision Conference , 2015. -
plain- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: unclear — It is unclear whether the paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
A result the paper states — it cites no source for it, so there is none to check. The quote shows it is in the paper.
In the paper
“Unlike the SITW dataset, both audio and video for each speaker is released.”
plain- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: unclear — It is unclear whether the paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
In the paper
“| Active speaker verification | [31] | 1.000 | 0.613 |”
What the sources say
- Not checked yetcites [31] [31] P. Chakravarty and T. Tuytelaars, 'Cross-modal supervision for learning active speaker detection in video,' arXiv preprint arXiv:1603.08907 , 2016. -
plain- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: unclear — It is unclear whether the paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
A result the paper states — it cites no source for it, so there is none to check. The quote shows it is in the paper.
In the paper
“reduces the number of parameters from 319M in VGG-M to 67M in our network”
plainstates a cause- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: unclear — It is unclear whether the paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
A result the paper reports from its own data — it cites no other source for it, so there is none to check. The quote shows it is in the paper.
In the paper
“| # of videos per POI | 36 / 18 / 8 |”
plain- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: unclear — It is unclear whether the paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
15 more claims
Create a free account to see all 50 claims and to verify the 16 references still open.
- Probably not a substantial claim. 9
- It may state two separate claims as one. 23
- The quote only partly states this claim. 5
Claims other papers make about this one
Each is checked against this paper’s own text: does it say what the citing paper says it does?
The cited source says what the claim attributes to it.
In the citing paper
“Meanwhile, the emergence of large-scale VSA datasets [28, 29, 30, 31, 32] promoted the further development of deep learning based VSA research.”
From Deep learning for visual speech analysis: A survey (2024) · cites this paper as [31]
In the citing paper
“Although it is mainly built for speaker identification, it also can be used for VSG.”
From Deep learning for visual speech analysis: A survey (2024) · cites this paper as [31]
In the citing paper
“It contains over 100k utterances from 1,251 celebrities.”
From Deep learning for visual speech analysis: A survey (2024) · cites this paper as [31]
In the citing paper
“VoxCeleb1 [31] is a large-scale text-independent audio-visual dataset collected from open-source YouTube media.”
From Deep learning for visual speech analysis: A survey (2024) · cites this paper as [31]
Citations
Cites
Deep Residual Learning for Image Recognition
He, K., Zhang, X., Ren, S., & Sun, J. (2016) · 2016 IEEE Conference on Computer Vision and Pattern Recognition (CVPR) · doi:10.1109/cvpr.2016.90 · Cited by 235,474
Batch Normalization: Accelerating Deep Network Training by Reducing Internal Covariate Shift
Ioffe, S., & Szegedy, C. (2015) · arXiv · doi:10.48550/arxiv.1502.03167 · Cited by 27,818 · cited 1×
“The network is trained using batch normalisation [40] and all hyper-parameters (e.”
MethodBatch Normalization: Accelerating Deep Network Training by Reducing Internal Covariate Shift
Ioffe, S. (2024) · arXiv (Cornell University) · doi:10.57702/o9raffed · Cited by 14,214
Deep Residual Learning for Image Recognition
He, K., Zhang, X., Ren, S., & Sun, J. (2015) · arXiv · doi:10.48550/arxiv.1512.03385 · Cited by 4,196
One millisecond face alignment with an ensemble of regression trees
Kazemi, V., & Sullivan, J. (2014) · 2014 IEEE Conference on Computer Vision and Pattern Recognition · doi:10.1109/cvpr.2014.241 · Cited by 2,831 · cited 1×
“Facial landmark positions are detected for each face detection using the regression tree based method of [33].”
MethodCNN architectures for large-scale audio classification
Hershey, S., Chaudhuri, S., Ellis, D. P. W., Gemmeke, J. F., Jansen, A., Moore, R. C., Plakal, M., Platt, D., Saurous, R. A., Seybold, B., et al. (2017) · 2017 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP) · doi:10.1109/icassp.2017.7952132 · Cited by 2,023
MatConvNet - Convolutional Neural Networks for MATLAB
Vedaldi, A., & Lenc, K. (2014) · arXiv (Cornell University) · doi:10.48550/arxiv.1412.4564 · Cited by 578 · cited 1×
“Our implementation is based on the deep learning toolbox MatConvNet [39] and trained on a NVIDIA TITAN X GPU.”
MethodCNN Architectures for Large-Scale Audio ClassificationPreprint
Hershey, S., Chaudhuri, S., Ellis, D. P. W., Gemmeke, J. F., Jansen, A., Moore, R. C., Plakal, M., Platt, D., Saurous, R. A., Seybold, B., et al. (2016) · arXiv (Cornell University) · doi:10.48550/arxiv.1609.09430 · Cited by 217 · cited 1×
“This has led to a very recent shift from hand-crafted features to the domain of deep CNNs which can be applied to higher dimensional inputs [13, 14] and for speaker iden-tification [15].”
BackgroundCross-Modal Supervision for Learning Active Speaker Detection in Video
Chakravarty, P., & Tuytelaars, T. (2016) · Lecture notes in computer science · doi:10.1007/978-3-319-46454-1_18 · Cited by 70 · cited 2×
“Precision-recall curves for both tasks on their respective benchmark datasets [30, 31] are shown in Figure 1 right, and the values at the operating point are given in Table 3.”
BackgroundCross-modal Supervision for Learning Active Speaker Detection in VideoPreprint
Chakravarty, P., & Tuytelaars, T. (2016) · arXiv (Cornell University) · doi:10.48550/arxiv.1603.08907 · Cited by 1 · cited 2×
“Precision-recall curves for both tasks on their respective benchmark datasets [30, 31] are shown in Figure 1 right, and the values at the operating point are given in Table 3.”
Background
Research works printed but not linked: 28
- [2][2] A. Krizhevsky, I. Sutskever, and G. E. Hinton, 'ImageNet classification with deep convolutional neural networks,' in Advances in Neural Information Processing Systems , pp. 1106-1114, 2012. -
- [5][5] M. McLaren, L. Ferrer, D. Castan, and A. Lawson, 'The speakers in the wild (SITW) speaker recognition database,' INTERSPEECH , vol. 2016, 2016. -
- [6][6] O. Russakovsky, J. Deng, H. Su, J. Krause, S. Satheesh, S. Ma, S. Huang, A. Karpathy, A. Khosla, M. Bernstein, A. Berg, and F. Li, 'Imagenet large scale visual recognition challenge,' International Journal of Computer Vision , 2015. -
- [7][7] D. A. Reynolds, T. F. Quatieri, and R. B. Dunn, 'Speaker verification using adapted gaussian mixture models,' Digital signal processing , vol. 10, no. 1-3, pp. 19-41, 2000. -
- [8][8] D. A. Reynolds and R. C. Rose, 'Robust text-independent speaker identification using gaussian mixture speaker models,' IEEE transactions on speech and audio processing , vol. 3, no. 1, pp. 7283, 1995. -
- [10][10] N. Dehak, P. J. Kenny, R. Dehak, P. Dumouchel, and P. Ouellet, 'Front-end factor analysis for speaker verification,' IEEE Transactions on Audio, Speech, and Language Processing , vol. 19, no. 4, pp. 788-798, 2011. -
- [11][11] U. H. Yapanel, X. Zhang, and J. H. Hansen, 'High performance digit recognition in real car environments.,' in INTERSPEECH , 2002. -
- [12][12] J. H. Hansen, R. Sarikaya, U. H. Yapanel, and B. L. Pellom, 'Robust speech recognition in noise: an evaluation using the spine corpus.,' in INTERSPEECH , pp. 905-908, 2001. -
- [13][13] T. N. Sainath, R. J. Weiss, A. W. Senior, K. W. Wilson, and O. Vinyals, 'Learning the speech front-end with raw waveform CLDNNs,' in INTERSPEECH , pp. 1-5, 2015. -
- [15][15] Y. Lukic, C. Vogt, O. D¨ urr, and T. Stadelmann, 'Speaker identification and clustering using convolutional neural networks,' in IEEE 26th International Workshop on Machine Learning for Signal Processing (MLSP) , pp. 1-6, IEEE, 2016. -
- [16][16] D. van der Vloed, J. Bouten, and D. A. van Leeuwen, 'NFIFRITS: a forensic speaker recognition database and some first experiments,' in The Speaker and Language Recognition Workshop , 2014. -
- [17][17] J. Hennebert, H. Melin, D. Petrovska, and D. Genoud, 'POLYCOST: a telephone-speech database for speaker recognition,' Speech communication , vol. 31, no. 2, pp. 265-270, 2000. -
- [18][18] J. B. Millar, J. P. Vonwiller, J. M. Harrington, and P. J. Dermody, 'The Australian national database of spoken language,' in Proceedings of the IEEE International Conference on Acoustics, Speech and Signal Processing , vol. 1, pp. I-97, IEEE, 1994. -
- [19][19] J. S. Garofolo, L. F. Lamel, W. M. Fisher, J. G. Fiscus, and D. S. Pallett, 'DARPA TIMIT acoustic-phonetic continous speech corpus CD-ROM. NIST speech disc 1-1.1,' NASA STI/Recon technical report , vol. 93, 1993. -
- [21][21] R. Woo, A. Park, and T. J. Hazen, 'The MIT Mobile Device Speaker Verification Corpus: Data collection and preliminary experiments,' The Speaker and Language Recognition Workshop , 2006. -
- [23][23] A. Janin, D. Baron, J. Edwards, D. Ellis, D. Gelbart, N. Morgan, B. Peskin, T. Pfau, E. Shriberg, A. Stolcke, et al. , 'The ICSI meeting corpus,' in Proceedings of the IEEE International Conference on Acoustics, Speech and Signal Processing , vol. 1, IEEE, 2003. -
- [24][24] I. McCowan, J. Carletta, W. Kraaij, S. Ashby, S. Bourban, M. Flynn, M. Guillemot, T. Hain, J. Kadlec, V. Karaiskos, et al. , 'The AMI meeting corpus,' in International Conference on Methods and Techniques in Behavioral Research , vol. 88, 2005. -
- [25][25] P. Bell, M. J. Gales, T. Hain, J. Kilgour, P. Lanchantin, X. Liu, A. McParland, S. Renals, O. Saz, M. Wester, et al. , 'The MGB challenge: Evaluating multi-genre broadcast media recognition,' in IEEE Workshop on Automatic Speech Recognition and Understanding , pp. 687-693, IEEE, 2015. -
- [27][27] J. J. Godfrey, E. C. Holliman, and J. McDaniel, 'Switchboard: Telephone speech corpus for research and development,' in Proceedings of the IEEE International Conference on Acoustics, Speech and Signal Processing , vol. 1, pp. 517-520, IEEE, 1992. -
- [28][28] W. M. Fisher, G. R. Doddington, and K. M. Goudie-Marshall, 'The DARPA speech recognition research database: specifications and status,' in Proc. DARPA Workshop on speech recognition , pp. 93-99, 1986. -
- [30][30] O. M. Parkhi, A. Vedaldi, and A. Zisserman, 'Deep face recognition,' in Proceedings of the British Machine Vision Conference , 2015. -
- [32][32] D. E. King, 'Dlib-ml: A machine learning toolkit,' The Journal of Machine Learning Research , vol. 10, pp. 1755-1758, 2009. -
- [34][34] J. S. Chung and A. Zisserman, 'Lip reading in the wild,' in Proceedings of the Asian Conference on Computer Vision , 2016. -
- [35][35] M. Everingham, J. Sivic, and A. Zisserman, 'Taking the bite out of automatic naming of characters in TV video,' Image and Vision Computing , vol. 27, no. 5, 2009. -
- [36][36] J. S. Chung and A. Zisserman, 'Out of time: automated lip sync in the wild,' in Workshop on Multi-view Lip-reading, ACCV , 2016. -
- [37][37] K. Chatfield, K. Simonyan, A. Vedaldi, and A. Zisserman, 'Return of the devil in the details: Delving deep into convolutional nets,' in Proceedings of the British Machine Vision Conference , 2014. -
- [38][38] S. Chopra, R. Hadsell, and Y. LeCun, 'Learning a similarity metric discriminatively, with application to face verification,' in Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition , vol. 1, pp. 539-546, IEEE, 2005. -
- [41][41] S. Ioffe, 'Probabilistic linear discriminant analysis,' in Proceedings of the European Conference on Computer Vision , pp. 531542, Springer, 2006.
Cited by
Items 201–300 of 1,008
MAAS: Multi-modal Assignation for Active Speaker Detection
Alcázar, J. L., Heilbron, F. C., Thabet, A., & Ghanem, B. (2021) · IEEE/CVF International Conference on Computer Vision (ICCV) · doi:10.1109/iccv48922.2021.00033 · Cited by 75
“[6] presented an improved architecture of their previous work [10], which trains a large 3D model with the need for largescale audiovisual pre-training [35].”
BackgroundTowards Learning Universal Audio Representations
Wang, L., Luc, P., Wu, Y., Recasens, A., Smaira, L., Brock, A., Jaegle, A., Alayrac, J.-B., Dieleman, S., Carreira, J., et al. (2022) · IEEE International Conference on Acoustics Speech and Signal Processing · doi:10.1109/icassp43922.2022.9746790 · Cited by 101
Contrastive Self-Supervised Learning for Text-Independent Speaker Verification
Zhang, H., Zou, Y., & Wang, H. (2021) · IEEE International Conference on Acoustics Speech and Signal Processing · doi:10.1109/icassp39728.2021.9413351 · Cited by 51
“We use Voxceleb1 [20] and Voxceleb2 [21] datasets to evaluate the effectiveness of the speaker verification sys-tem.”
MethodTowards Understanding and Mitigating Audio Adversarial Examples for Speaker Recognition
Chen, G., Zhao, Z., Song, F., Chen, S., Fan, L., Wang, F., & Wang, J. (2022) · IEEE Transactions on Dependable and Secure Computing · doi:10.1109/tdsc.2022.3220673 · Cited by 63
Hong, F.-T., & Xu, D. (2023) · IEEE/CVF International Conference on Computer Vision (ICCV) · doi:10.1109/iccv51070.2023.02108 · Cited by 98
PuppeteerGAN: Arbitrary Portrait Animation With Semantic-Aware Appearance Transformation
Chen, Z. Q., Wang, C., Yuan, B., & Tao, D. (2020) · IEEE Computer Society Conference on Computer Vision and Pattern Recognition/Proceedings - IEEE Computer Society Conference on Computer Vision and Pattern Recognition/Proceedings · doi:10.1109/cvpr42600.2020.01353 · Cited by 62
“In comparison experiments, our model was trained on a subset of VoxCeleb1 [24] with video sequences of 672 person for training and 50 for test.”
InfluentialDual-Generator Face Reenactment
Hsu, G.-S., Tsai, C.-H., & Wu, H.-Y. (2022) · 2022 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR) · doi:10.1109/cvpr52688.2022.00072 · Cited by 41
“VoxCeleb1 The VoxCeleb1 dataset [16] contains over 100k utterances for 1,251 celebrities, extracted from the YouTube videos, and is divided into the training and testing sets.”
BackgroundMethodFace-Voice Matching using Cross-modal Embeddings
Horiguchi, S., Kanda, N., & Nagamatsu, K. (2018) · ACM International Conference on Multimedia (ACM MM) · doi:10.1145/3240508.3240601 · Cited by 62
“It is composed of samples from 1,078 persons, who are included in both the MS-Celeb-1M dataset [10] and the Vox-Celeb dataset [30].”
BackgroundHow to Improve Your Speaker Embeddings Extractor in Generic Toolkits
Zeinali, H., Burget, L., Rohdin, J., Stafylakis, T., & Černocký, J. (2019) · IEEE International Conference on Acoustics Speech and Signal Processing · doi:10.1109/icassp.2019.8683445 · Cited by 47
Within-Sample Variability-Invariant Loss for Robust Speaker Recognition Under Noisy Environments
Cai, D., Cai, W., & Li, M. (2020) · IEEE International Conference on Acoustics Speech and Signal Processing · doi:10.1109/icassp40776.2020.9053407 · Cited by 49
“The experiments are conducted on Voxceleb 1 dataset [25].”
MethodDelving into VoxCeleb: Environment Invariant Speaker Recognition
Chung, J. S., Huh, J., & Mun, S. (2020) · The Speaker and Language Recognition Workshop (Odyssey) · doi:10.21437/odyssey.2020-49 · Cited by 64
Self-citationA Multi-View Approach to Audio-Visual Speaker Verification
Sarı, L., Singh, K., Zhou, J., Torresani, L., Singhal, N., & Saraf, Y. (2021) · IEEE International Conference on Acoustics Speech and Signal Processing · doi:10.1109/icassp39728.2021.9414260 · Cited by 60
“To our knowledge, we achieve the best reported AV speaker verification performance on the VoxCeleb1 [11] and VoxCeleb2 [12] datasets.”
InfluentialVoxSRC 2019: The first VoxCeleb Speaker Recognition ChallengePreprint
Chung, J. S., Nagrani, A., Coto, E., Xie, W., McLaren, M., Reynolds, D. A., & Zisserman, A. (2019) · arXiv (Cornell University) · doi:10.48550/arxiv.1912.02522 · Cited by 47
“The VoxCeleb datasets [4–6] consist of speech segments from unconstrained YouTube videos for several thousand individuals.”
MethodSelf-citationThe ins and outs of speaker recognition: lessons from VoxSRC 2020
Kwon, Y., Heo, H.-S., Lee, B., & Chung, J. S. (2021) · IEEE International Conference on Acoustics Speech and Signal Processing · doi:10.1109/icassp39728.2021.9413948 · Cited by 75
Self-citationShon, S., Oh, T.-H., & Glass, J. R. (2019) · IEEE International Conference on Acoustics Speech and Signal Processing · doi:10.1109/icassp.2019.8683477 · Cited by 46
M3F: Multi-Modal Continuous Valence-Arousal Estimation in the Wild
Zhang, Y., Huang, R., Zeng, J., & Shan, S. (2020) · International Conference on Automatic Face and Gesture Recognition · doi:10.1109/fg47880.2020.00098 · Cited by 47
“For AU features, we use the self-supervised 256-dimensional encoder features using TCAE [14], which is trained on the union of VoxCeleb1 [15] and VoxCeleb2 [10] datasets.”
MethodDeepFake Disrupter: The Detector of DeepFake Is My Friend
Wang, X., Huang, J., Ma, S., Nepal, S., & Xu, C. (2022) · 2022 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR) · doi:10.1109/cvpr52688.2022.01450 · Cited by 45
“Experimental results on CelebA [13] and VoxCeleb1 [16] datasets demonstrate that the proposed DeepFake disrupter can effectively protect original real images/videos from being used as a source for making DeepFake data.”
MethodOn Joint Optimization of Automatic Speaker Verification and Anti-Spoofing in the Embedding Space
Gomez-Alanis, A., Gonzalez-Lopez, J. A., Dubagunta, S. P., Peinado, A. M., & Magimai.-Doss, M. (2021) · IEEE Transactions on Information Forensics and Security · doi:10.1109/tifs.2020.3039045 · Cited by 45
“Moreover, the development set of the Voxceleb1 [59] database, which includes a total of 1,231 training speakers, was combined with the bonafide training sets of the ASVspoof 2019 database in order to train the PLDA and b-vector ASV scoring systems.”
MethodInfluentialCompact Temporal Trajectory Representation for Talking Face Video Compression
Chen, B., Wang, Z., Li, B., Wang, S., & Ye, Y. (2023) · IEEE Transactions on Circuits and Systems for Video Technology · doi:10.1109/tcsvt.2023.3271130 · Cited by 58
“The 40 testing sequences selected and pre-processed from 300-VW dataset [71] and VoxCeleb dataset [70]. feature representation.”
MethodHyperReenact: One-Shot Reenactment via Jointly Learning to Refine and Retarget Faces
Bounareli, S., Tzelepis, C., Argyriou, V., Patras, I., & Tzimiropoulos, G. (2023) · IEEE/CVF International Conference on Computer Vision (ICCV) · doi:10.1109/iccv51070.2023.00657 · Cited by 80
Islam, N., & Shin, S. (2024) · IEEE Open Journal of the Communications Society · doi:10.1109/ojcoms.2024.3425314 · Cited by 64
“Future efforts could focus on refining the FL to enhance scalability and robustness, ensuring consistent performance across a wide range of network conditions and participant variability.”
BackgroundSelf-Supervised Speaker Embeddings
Stafylakis, T., Rohdin, J., Plchot, O., Mizera, P., & Burget, L. (2019) · Interspeech (USB) · doi:10.21437/interspeech.2019-2842 · Cited by 58
Deep Latent Space Learning for Cross-Modal Mapping of Audio and Visual Signals
Nawaz, S., Janjua, M. K., Gallo, I., Mahmood, A., & Calefati, A. (2019) · International Conference on Digital Image Computing: Techniques and Applications (DICTA) · doi:10.1109/dicta47822.2019.8945863 · Cited by 65
“Without characterizing the background noise and variation, holistic understanding and robust learning is not possible [6].”
BackgroundMethodInfluentialJhu-HLTCOE System for the Voxsrc Speaker Recognition Challenge
Garcia-Romero, D., McCree, A. V., Snyder, D., & Sell, G. (2020) · IEEE International Conference on Acoustics Speech and Signal Processing · doi:10.1109/icassp40776.2020.9053209 · Cited by 45
“We used the VoxCeleb1-E (cleaned version) as our validation set [29, 30].”
MethodOptimization of data-driven filterbank for automatic speaker verification
Sarangi, S. K., Sahidullah, M., & Saha, G. (2020) · Digital Signal Processing · doi:10.1016/j.dsp.2020.102795 · Cited by 58
“The values of Cmiss, Cfa and Ptar are chosen according to the evaluation plan of the respective corpus [7, 65, 66] and their values are shown in Table 5.”
MethodVoice In Ear: Spoofing-Resistant and Passphrase-Independent Body Sound Authentication
Gao, Y., Jin, Y., Chauhan, J., Choi, S., Li, J., & Jin, Z. (2021) · Proceedings of the ACM on Interactive Mobile Wearable and Ubiquitous Technologies · doi:10.1145/3448113 · Cited by 49
“We first selected a sub-set which contains 100 subjects of the VoxCeleb dataset [34] (i.”
MethodAutomated Recognition of Alzheimer’s Dementia Using Bag-of-Deep-Features and Model Ensembling
Syed, Z. S., Syed, M. S. S., Lech, M., & Pirogova, E. (2021) · IEEE Access · doi:10.1109/access.2021.3090321 · Cited by 64
“For the audio modality, they used x-vector and i-vector based acoustic embeddings which were previously trained on SRE [34] and VoxCeleb [35] datasets.”
MethodShah, J., Singla, Y. K., Chen, C., & Shah, R. R. (2021) · arXiv (Cornell University) · doi:10.48550/arxiv.2101.00387 · Cited by 89
“To this end, we perform two representative tasks: speaker recognition on Voxceleb [46] (which uses audio features primarily), and phone classification on LibriSpeech (which uses pronunciation features).”
MethodDeep neural networks for automatic speech processing: a survey from large corpora to limited data
Roger, V., Farinas, J., & Pinquier, J. (2022) · EURASIP Journal on Audio Speech and Music Processing · doi:10.1186/s13636-022-00251-w · Cited by 66
Ensemble Additive Margin Softmax for Speaker Verification
Yu, Y.-Q., Fan, L., & Li, W.-J. (2019) · IEEE International Conference on Acoustics Speech and Signal Processing · doi:10.1109/icassp.2019.8683649 · Cited by 46
“In particular, the development set of VoxCeleb2 is used for training and the test set of VoxCeleb1 is used for testing.”
BackgroundMethodInfluentialLow Bandwidth Video-Chat Compression using Deep Generative Models
Oquab, M., Stock, P., Gafni, O., Haziza, D., Xu, T., Zhang, P., Çelebi, O., Hasson, Y., Labatut, P., Bose-Kolanu, B., et al. (2021) · IEEE/CVF Conference on Computer Vision and Pattern Recognition Workshops (CVPRW) · doi:10.1109/cvprw53098.2021.00271 · Cited by 59
“Our mobile models are trained on the DFDC dataset [13] rather than the VoxCeleb [26] dataset, in contrast to the original work of [33] (though we provide evaluation numbers for comparison and reference).”
MethodInfluentialSelective Listening by Synchronizing Speech With Lips
Pan, Z., Tao, R., Xu, C., & Li, H. (2022) · IEEE/ACM Transactions on Audio Speech and Language Processing · doi:10.1109/taslp.2022.3153258 · Cited by 70
“using a face detector as in [95], and a tracker that is based on ROI overlap [96].”
MethodFrom Speaker Verification to Multispeaker Speech Synthesis, Deep Transfer with Feedback Constraint
Cai, Z., Zhang, C., & Li, M. (2020) · Interspeech (USB) · doi:10.21437/interspeech.2020-1032 · Cited by 54
Voice Activity Detection in the Wild: A Data-Driven Approach Using Teacher-Student Training
Dinkel, H., Wang, S., Xu, X., Wu, M., & Yu, K. (2021) · IEEE/ACM Transactions on Audio Speech and Language Processing · doi:10.1109/taslp.2021.3073596 · Cited by 59
“These three datasets are: Voxceleb1 ( V 1 ) [36], VoxCeleb2 ( V 2 ) [37], [38], as well as V 3 which is a combination of the SRE datasets [39] and Switchboard datasets [40].”
MethodAugmentation adversarial training for unsupervised speaker recognition.Preprint
Huh, J., Heo, H. S., Kang, J., Watanabe, S., & Chung, J. S. (2020) · arXiv (Cornell University) · Cited by 52
Self-citationMulti-View Self-Attention Based Transformer for Speaker Recognition
Wang, R., Ao, J., Zhou, L., Liu, S., Wei, Z., Ko, T., Li, Q., & Zhang, Y. (2022) · IEEE International Conference on Acoustics Speech and Signal Processing · doi:10.1109/icassp43922.2022.9746639 · Cited by 57
Deep Person Generation: A Survey from the Perspective of Face, Pose, and Cloth Synthesis
Sha, T., Zhang, W., Shen, T., Li, Z., & Mei, T. (2022) · ACM Computing Surveys · doi:10.1145/3575656 · Cited by 55
“ObamaSet 2017 [150] 14 1 — √ × VoxCeleb 2017 [117] 352 1.”
Deep speaker embeddings for Speaker Verification: Review and experimental comparison
Jakubec, M., Jarina, R., Lieskovská, E., & Kasák, P. (2023) · Engineering Applications of Artificial Intelligence · doi:10.1016/j.engappai.2023.107232 · Cited by 62
Speech driven video editing via an audio-conditioned diffusion model
Bigioi, D., Basak, S., Stypułkowski, M., Zięba, M., Jordan, H., McDonnell, R., & Corcoran, P. (2024) · Image and Vision Computing · doi:10.1016/j.imavis.2024.104911 · Cited by 62
Voice Mimicry Attacks Assisted by Automatic Speaker Verification
Vestman, V., Kinnunen, T., Hautamäki, R. G., & Sahidullah, M. (2019) · Computer Speech & Language · doi:10.1016/j.csl.2019.05.005 · Cited by 49
“The attacker’s ASV is used as a voice search tool to find the closest speakers from the combination of VoxCeleb1 [20] and Voxceleb2 [21] to each of the locally recruited subjects (described in Section 5).”
MethodInvestigation of Specaugment for Deep Speaker Embedding Learning
Wang, S., Rohdin, J., Plchot, O., Burget, L., Yu, K., & Černocký, J. (2020) · IEEE International Conference on Acoustics Speech and Signal Processing · doi:10.1109/icassp40776.2020.9053481 · Cited by 48
“In the last few years, the state-of-the-art embeddings for speaker recognition have shifted from the so-called i-vectors[1] produced by generative shallow models towards embeddings produced by discriminatively trained deep neural networks (DNN) [2, 3, 4, 5, 6, 7, 8, 9].”
BackgroundAlignNet: A Unifying Approach to Audio-Visual Alignment
Wang, J., Fang, Z., & Zhao, H. (2020) · IEEE Winter Conference on Applications of Computer Vision · doi:10.1109/wacv45572.2020.9093345 · Cited by 52
TRILLsson: Distilled Universal Paralinguistic Speech Representations
Shor, J., & Venugopalan, S. (2022) · Interspeech (USB) · doi:10.21437/interspeech.2022-118 · Cited by 73
“VoxCeleb∗ [18] Speaker ID 1,251 12,052 8.”
Improving Zero-shot Voice Style Transfer via Disentangled Representation LearningPreprint
Yuan, S., Cheng, P., Zhang, R., Hao, W., Gan, Z., & Carin, L. (2021) · arXiv (Cornell University) · doi:10.48550/arxiv.2103.09420 · Cited by 61
“Speaker encoder pretraining Our speaker encoder is pretrained with GE2E loss on a combination of VoxCeleb1 (Nagrani et al., 2017) and Librispeech (Panayotov et al., 2015) datasets, in total of 3549 speakers.”
MethodZolnoori, M., Zolnour, A., & Topaz, M. (2023) · Artificial Intelligence in Medicine · doi:10.1016/j.artmed.2023.102624 · Cited by 60
Investigation of End-to-End Speaker-Attributed ASR for Continuous Multi-Talker Recordings
Kanda, N., Chang, X., Gaur, Y., Wang, X., Meng, Z., Chen, Z., & Yoshioka, T. (2021) · IEEE Spoken Language Technology Workshop (SLT) · doi:10.1109/slt48900.2021.9383600 · Cited by 62
“For the speaker profile, we used a 128-dim d-vector [31], whose extractor was separately trained on VoxCeleb Corpus [39, 40].”
MethodPrivacy and Utility of X-Vector Based Speaker Anonymization
Srivastava, B. M. L., Maouche, M., Sahidullah, M., Vincent, E., Bellet, A., Tommasi, M., Tomashenko, N., Wang, X., & Yamagishi, J. (2022) · IEEE/ACM Transactions on Audio Speech and Language Processing · doi:10.1109/taslp.2022.3190741 · Cited by 68
“Among the components of the anonymization system, the ASR AM is trained on the trainclean-100 and train-other-500 subsets of LibriSpeech [33], the x-vector extractor is trained on VoxCeleb1 [34] and VoxCeleb2 [35], and the speech synthesis AM and NSF model are trained on the train-clean-100 subset of LibriTTS [36].”
MethodKanda, N., Xiao, X., Gaur, Y., Wang, X., Meng, Z., Chen, Z., & Yoshioka, T. (2022) · IEEE International Conference on Acoustics Speech and Signal Processing · doi:10.1109/icassp43922.2022.9746225 · Cited by 41
MMCosine: Multi-Modal Cosine Loss Towards Balanced Audio-Visual Fine-Grained Learning
Xu, R., Feng, R., Zhang, S.-X., & Hu, D. (2023) · IEEE International Conference on Acoustics Speech and Signal Processing · doi:10.1109/icassp49357.2023.10096655 · Cited by 82
“Following previous study [20, 21, 22], we train our model on Voxceleb2 [18] dev partition and validate on Voxceleb1 [17] and the trial list randomly sampled from Voxceleb2 test.”
MethodFace Animation with an Attribute-Guided Diffusion Model
Zeng, B., Liu, X., Gao, S., Liu, B., Li, H., Liu, J., & Zhang, B. (2023) · IEEE/CVF Conference on Computer Vision and Pattern Recognition Workshops (CVPRW) · doi:10.1109/cvprw59228.2023.00070 · Cited by 78
“Quantitative comparison of same-identity reconstruction on VoxCeleb [35].”
BackgroundMethodSelf-Supervised Learning With Cluster-Aware-DINO for High-Performance Robust Speaker Verification
Han, B., Chen, Z., & Qian, Y. (2023) · IEEE/ACM Transactions on Audio Speech and Language Processing · doi:10.1109/taslp.2023.3331949 · Cited by 54
“The experiments on Voxceleb showed that our newly proposed CA-DINO with DLG-LC is superior and achieves the new state-of-the-art performance for self-supervised speaker ver-ification.”
MethodInfluentialDisentangling Style Factors from Speaker Representations
Williams, J. A., & King, S. (2019) · Interspeech (USB) · doi:10.21437/interspeech.2019-1769 · Cited by 56
“That corpus contains approximately 2,000 unique celebrity speakers [25, 26] and was augmented with noise during training.”
BackgroundJung, Y., Kye, S. M., Choi, Y., Jung, M., & Kim, H. (2020) · Interspeech (USB) · doi:10.21437/interspeech.2020-1025 · Cited by 41
“The dev set of VoxCeleb2 is used for training and the test set of VoxCeleb1 is used for testing.”
BackgroundMethodInfluentialInvestigating Robustness of Adversarial Samples Detection for Automatic Speaker Verification
Xu, L., Li, N., Zhong, J., Wu, X., Liu, X., Su, D., Yu, D., & Meng, H. M. L. (2020) · Interspeech (USB) · doi:10.21437/interspeech.2020-2441 · Cited by 50
Move2Hear: Active Audio-Visual Source Separation
Majumder, S., Al-Halah, Z., & Grauman, K. L. (2021) · IEEE/CVF International Conference on Computer Vision (ICCV) · doi:10.1109/iccv48922.2021.00034 · Cited by 66
Deep speaker conditioning for speech emotion recognition
Triantafyllopoulos, A., Liu, S., & Schuller, B. W. (2021) · IEEE International Conference on Multimedia and Expo (ICME) · doi:10.1109/icme51207.2021.9428217 · Cited by 49
“This methodology is quite common in the speaker identification literature [22, 23].”
Method井上, 中., & Goto, K. (2020) · arXiv (Cornell University) · doi:10.48550/arxiv.2006.04326 · Cited by 42
“Experimental Settings We used the VoxCeleb dataset [23], [24] for evaluating our proposed framework.”
MethodFarsiani, S., Izadkhah, H., & Lotfi, S. (2022) · Computers & Electrical Engineering · doi:10.1016/j.compeleceng.2022.107882 · Cited by 39
Multi-Resolution Multi-Head Attention in Deep Speaker Embedding
Wang, Z., Yao, K., Li, X., & Fang, S. (2020) · IEEE International Conference on Acoustics Speech and Signal Processing · doi:10.1109/icassp40776.2020.9053217 · Cited by 41
“The challenging VoxCeleb1 verification task [15] Network architecture.”
BackgroundAn Iterative Framework for Self-Supervised Deep Speaker Representation Learning
Cai, D., Wang, W., & Li, M. (2021) · IEEE International Conference on Acoustics Speech and Signal Processing · doi:10.1109/icassp39728.2021.9414713 · Cited by 56
“For evaluation, the development set and test set of Voxceleb 1 are used [23].”
MethodMetaAudio: A Few-Shot Audio Classification Benchmark
Heggan, C., Budgett, S., Hospedales, T. M., & Yaghoobi, M. (2022) · Lecture notes in computer science · doi:10.1007/978-3-031-15919-0_19 · Cited by 60
Salvati, D., Drioli, C., & Foresti, G. L. (2023) · Expert Systems with Applications · doi:10.1016/j.eswa.2023.119750 · Cited by 39
Peng, J., Plchot, O., Stafylakis, T., Mošner, L., Burget, L., & Černocký, J. (2023) · IEEE Spoken Language Technology Workshop (SLT) · doi:10.1109/slt54892.2023.10022775 · Cited by 82
“The SV performance is evaluated on the VoxCeleb corpus [23, 24], which is a widely used large-scale text-independent speaker verification data-set.”
Self-Supervised Audio Teacher-Student Transformer for Both Clip-Level and Frame-Level Tasks
Li, X. W., Shao, N., & Li, X. (2024) · IEEE/ACM Transactions on Audio Speech and Language Processing · doi:10.1109/taslp.2024.3352248 · Cited by 88
“We use the Voxceleb1 dataset [39], with 1,251 speakers.”
MethodTalking Head Generation with Probabilistic Audio-to-Visual Diffusion Priors
Yu, Z., Yin, Z., Zhou, D., Wang, D., Wong, F., & Wang, B. (2023) · IEEE/CVF International Conference on Computer Vision (ICCV) · doi:10.1109/iccv51070.2023.00703 · Cited by 77
“Practically, we found it easy for P a 2 nl to quickly learn how to fill in the masked region using hints from the non-lip features in the unmasked regions, generating continuous feature prediction without any extra design in training loss.”
MethodNPLDA: A Deep Neural PLDA Model for Speaker Verification
Ramoji, S., Krishnan, P., & Ganapathy, S. (2020) · The Speaker and Language Recognition Workshop (Odyssey) · doi:10.21437/odyssey.2020-29 · Cited by 43
Jung, J.-W., Heo, H.-S., Shim, H.-J., & Yu, H.-J. (2019) · IEEE Workshop on Automatic Speech Recognition and Understanding (ASRU) · doi:10.1109/asru46091.2019.9004029 · Cited by 39
“In all the experiments described herein, we used VoxCeleb1 dataset, which comprises approximately 330 hours of audio of 1,251 speakers, at a sampling rate of 16 kHz [26].”
MethodNammous, M. K., Saeed, K., & Kobojek, P. (2020) · Journal of King Saud University - Computer and Information Sciences · doi:10.1016/j.jksuci.2020.03.011 · Cited by 38
“Later research (Yadav and Rai 2018) kept almost these portions whilst reusing the same VoxCeleb1 dataset, with 95.5% of the dataset dedicated for training and only 4.5% for testing.”
InfluentialChung, S.-W., Kang, H.-G., & Chung, J. S. (2020) · Interspeech (USB) · doi:10.21437/interspeech.2020-1113 · Cited by 48
Self-citationLearning Representations for Facial Actions From Unlabeled Videos
Li, Y., Zeng, J., & Shan, S. (2020) · IEEE Transactions on Pattern Analysis and Machine Intelligence · doi:10.1109/tpami.2020.3011063 · Cited by 51
The INESC-ID Multi-Modal System for the ADReSS 2020 Challenge
Pompili, A., Rolland, T., & Abad, A. (2020) · Interspeech (USB) · doi:10.21437/interspeech.2020-2833 · Cited by 46
QASR: QCRI Aljazeera Speech Resource A Large Scale Annotated Arabic Speech Corpus
Mubarak, H., Hussein, A., Chowdhury, S. A., & Ali, A. M. A. (2021) · doi:10.18653/v1/2021.acl-long.177 · Cited by 43
“The model is trained on Voxceleb1 (Nagrani et al., 2017) development set (containing 1 , 211 speakers and ≈ 147 K utterances).”
MethodEnd-to-End Speaker Verification via Curriculum Bipartite Ranking Weighted Binary Cross-Entropy
Bai, Z., Wang, J., Zhang, X.-L., & Chen, J. (2022) · IEEE/ACM Transactions on Audio Speech and Language Processing · doi:10.1109/taslp.2022.3161155 · Cited by 44
Structure-Aware Motion Transfer with Deformable Anchor Model
Tao, J., Wang, B., Xu, B., Ge, T., Jiang, Y., Li, W., & Duan, L. (2022) · IEEE Computer Society Conference on Computer Vision and Pattern Recognition/Proceedings - IEEE Computer Society Conference on Computer Vision and Pattern Recognition/Proceedings · doi:10.1109/cvpr52688.2022.00362 · Cited by 63
“• VoxCeleb1 [16] is a talking head dataset, containing 19522 training 525 test videos.”
BackgroundAnonymizing Speech with Generative Adversarial Networks to Preserve Speaker Privacy
Meyer, S., Tilli, P., Denisov, P., Lux, F., Koch, J., & Vu, N. T. (2023) · IEEE Spoken Language Technology Workshop (SLT) · doi:10.1109/slt54892.2023.10022601 · Cited by 67
“Thus, we use the x-vector and ECAPATDNN embedding extractors trained by SpeechBrain [33] on the 2,800 h speaker verification corpus VoxCeleb 1 and 2 [38, 39, 40].”
MethodTarget-Speaker Voice Activity Detection Via Sequence-to-Sequence Prediction
Cheng, M., Wang, W., Zhang, Y., Qin, X., & Li, M. (2023) · IEEE International Conference on Acoustics Speech and Signal Processing · doi:10.1109/icassp49357.2023.10094752 · Cited by 41
Saritha, B., Laskar, M. A., Kirupakaran, A. M., Laskar, R. H., Choudhury, M., & Shome, N. (2023) · Circuits Systems and Signal Processing · doi:10.1007/s00034-023-02542-9 · Cited by 40
“The experimentation is carried out on Part 1 of the RSR2015 [20] database and Vox-celeb1 dataset [27].”
InfluentialMasked Modeling Duo: Towards a Universal Audio Pre-Training Framework
Niizumi, D., Takeuchi, D., Ohishi, Y., Harada, N., & Kashino, K. (2024) · IEEE/ACM Transactions on Audio Speech and Language Processing · doi:10.1109/taslp.2024.3389636 · Cited by 93
“T UNING R ESULTS W ITH 95% CI”
BackgroundFace-Adapter for Pre-trained Diffusion Models with Fine-Grained ID and Attribute Control
Han, Y., Zhu, J., He, K., Chen, X., Ge, Y., Li, W., Li, X., Zhang, J., Wang, C., & Liu, Y. (2024) · Lecture notes in computer science · doi:10.1007/978-3-031-72973-7_2 · Cited by 111
A comprehensive overview of Generative AI (GAI): Technologies, applications, and challenges
Zhang, Z., Zhang, J., Zhang, X., & Mai, W. (2025) · Neurocomputing · doi:10.1016/j.neucom.2025.129645 · Cited by 59
Realistic Face Reenactment via Self-Supervised Disentangling of Identity and Pose
Zeng, X., Pan, Y., Wang, M., Zhang, J., & Liu, Y. (2020) · Proceedings of the AAAI Conference on Artificial Intelligence · doi:10.1609/aaai.v34i07.6970 · Cited by 50
“We conduct our ablation study and comparisons on VoxCeleb1 dataset (Nagrani, Chung, and Zisserman 2017; Nagrani, Albanie, and Zisserman 2018).”
Voice-Indistinguishability: Protecting Voiceprint In Privacy-Preserving Speech Data Release
Han, Y., Li, S., Cao, Y., Ma, Q., & Yoshikawa, M. (2020) · IEEE International Conference on Multimedia and Expo (ICME) · doi:10.1109/icme46284.2020.9102875 · Cited by 59
“We use the open-source x-vector system pre-trained on the augmented VoxCeleb-1 [25] and VoxCeleb-2 datasets [26].”
MethodDeep Normalization for Speaker Vectors
Cai, Y., Li, L., Abel, A. K., Zhu, X., & Wang, D. (2020) · IEEE/ACM Transactions on Audio Speech and Language Processing · doi:10.1109/taslp.2020.3039573 · Cited by 42
Revisiting the Statistics Pooling Layer in Deep Speaker Embedding Learning
Wang, S., Yang, Y., Qian, Y., & Yu, K. (2021) · International Symposium on Chinese Spoken Language Processing (ISCSLP) · doi:10.1109/iscslp49672.2021.9362097 · Cited by 41
“The x-vector style systems exhibits better performance on several popular datasets, including Vox-Celeb [10, 11] and NIST SRE.”
BackgroundAdversarial Defense for Automatic Speaker Verification by Cascaded Self-Supervised Learning Models
Wu, H., Li, X., Liu, A. T., Wu, Z., Meng, H. M. L., & Lee, H.-Y. (2021) · IEEE International Conference on Acoustics Speech and Signal Processing · doi:10.1109/icassp39728.2021.9413737 · Cited by 38
Seeking the Shape of Sound: An Adaptive Framework for Learning Voice-Face Association
Wen, P., Xu, Q., Jiang, Y., Yang, Z., He, Y., & Huang, Q. (2021) · IEEE Computer Society Conference on Computer Vision and Pattern Recognition/Proceedings - IEEE Computer Society Conference on Computer Vision and Pattern Recognition/Proceedings · doi:10.1109/cvpr46437.2021.01608 · Cited by 54
“Following previous work [30, 27], we evaluate the proposed method on a constructed dataset based on VoxCeleb [18] and VGGFace [20] datasets.”
MethodThe SpeakIn System for VoxCeleb Speaker Recognition Challange 2021Preprint
Zhao, M., Ma, Y., Liu, M., & Xu, M. (2021) · arXiv (Cornell University) · doi:10.48550/arxiv.2109.01989 · Cited by 74
“To evaluate the performance of our models, we used 5 test sets [1, 10] as our developing sets: • VoxCeleb1-O: 37,720 trails are sampled from the VoxCeleb1 test dataset with only 40 speakers.”
MethodSSAMBA: Self-Supervised Audio Representation Learning With Mamba State Space Model
Shams, S., Dindar, S. S., Jiang, X., & Mesgarani, N. (2024) · IEEE Spoken Language Technology Workshop (SLT) · doi:10.1109/slt61566.2024.10832304 · Cited by 71
“It includes speech from 1,251 speakers, with the task being to classify each utterance by its speaker identity [25].”
Who Said That?: Audio-Visual Speaker Diarisation of Real-World Meetings
Chung, J. S., Lee, B., & Han, I. (2019) · Interspeech (USB) · doi:10.21437/interspeech.2019-3116 · Cited by 48
Self-citationSpeaker Awareness for Speech Emotion Recognition
Assunção, G., Menezes, P. J., & Perdigão, F. (2020) · International Journal of Online and Biomedical Engineering (iJOE) · doi:10.3991/ijoe.v16i04.11870 · Cited by 42
“In accordance with the VGGVox model's implementation, and in order to take full advantage of all the provided audio, files were adapted to be 1 to 10 seconds in length as well.”
MethodInfluentialFakeBuster: A DeepFakes Detection Tool for Video Conferencing Scenarios
Mehta, V., Gupta, P., Subramanian, R., & Dhall, A. (2021) · 26th International Conference on Intelligent User Interfaces · doi:10.1145/3397482.3450726 · Cited by 34
“To get an equal number of real videos, we selected 4105 face videos from the pre-processed VoxCeleb [14] dataset.”
MethodInfluentialLearning to Aggregate and Personalize 3D Face from In-the-Wild Photo Collection
Zhang, Z., Ge, Y., Chen, R., Tai, Y. Y., Yan, Y., Yang, J. J., Wang, C., Li, J., & Huang, F. (2021) · IEEE Computer Society Conference on Computer Vision and Pattern Recognition/Proceedings - IEEE Computer Society Conference on Computer Vision and Pattern Recognition/Proceedings · doi:10.1109/cvpr46437.2021.01399 · Cited by 47
“A simple way is to use facial videos such as voxceleb [32] in constrained condition, but it may also have drawbacks: the quality of videos and various pose variations cannot be guaranteed.”
MethodOverlap-Aware Low-Latency Online Speaker Diarization Based on End-to-End Local Segmentation
Coria, J. M., Bredin, H., Ghannay, S., & Rosset, S. (2021) · IEEE Workshop on Automatic Speech Recognition and Understanding (ASRU) · doi:10.1109/asru51503.2021.9688044 · Cited by 38
Large-Scale Unsupervised Pre-Training for End-to-End Spoken Language Understanding
Wang, P., Wei, L., Cao, Y., Xie, J., & Nie, Z. (2020) · IEEE International Conference on Acoustics Speech and Signal Processing · doi:10.1109/icassp40776.2020.9053163 · Cited by 46
“We use more than 4000 hours of publicly available audio clips for the SLU component pretraining, including Common Voice, Libri Speech and VoxCeleb1&2 [25], as shown in Table 2.”
MethodRadha, K., Bansal, M., & Pachori, R. B. (2024) · Engineering Applications of Artificial Intelligence · doi:10.1016/j.engappai.2023.107661 · Cited by 43
Bhattacharya, G., Alam, J., & Kenny, P. J. (2019) · IEEE International Conference on Acoustics Speech and Signal Processing · doi:10.1109/icassp.2019.8682611 · Cited by 41
Speaker Characterization Using TDNN-LSTM Based Speaker Embedding
Chen, C.-P., Zhang, S.-Y., Yeh, C.-T., Wang, J., Wang, T., & Huang, C.-L. (2019) · IEEE International Conference on Acoustics Speech and Signal Processing · doi:10.1109/icassp.2019.8683185 · Cited by 36
“The SRE data is up-sampled to 16 kHz and combined with 16 kHz VoxCeleb to build speaker-embedding systems based on 16 kHz data.”
InfluentialKnowledge Distillation for Small Foot-print Deep Speaker Embedding
Wang, S., Yang, Y., Wang, T., Qian, Y., & Yu, K. (2019) · IEEE International Conference on Acoustics Speech and Signal Processing · doi:10.1109/icassp.2019.8683443 · Cited by 36
“045% EERs with PLDA and Cosine Distance Scoring, respectively, which is comparable to the results on the same dataset in the literature [14, 26, 27].”
BackgroundMethodResultInfluentialLü, L., Yu, J., Chen, Y. J., & Wang, Y. (2020) · Proceedings of the ACM on Interactive Mobile Wearable and Ubiquitous Technologies · doi:10.1145/3397320 · Cited by 49
“This is because the UBM is pre-trained using a large number of samples from different persons, e.g., trained with VoxCeleb dataset in this experiment, which enables UBM with the basic knowledge of individual uniqueness under the vocal tract.”
BackgroundMethodInfluentialAVA-AVD: Audio-visual Speaker Diarization in the Wild
Xu, Z., Song, Z., Tsutsui, S., Feng, C., Ye, M., & Shou, M. Z. (2022) · ACM International Conference on Multimedia (ACM MM) · doi:10.1145/3503161.3548027 · Cited by 73
Metadata
- Authors
- Nagrani, A., Chung, J. S., & Zisserman, A.
- Venue
- Interspeech 2017
- Publisher
- ISCA
- Published
- 2017-08-20
- Pages
- 2616-2620
- ISSN
- 2308-457X (ISSN-L)
- OpenAlex
- W2726515241
- Cited by
- 2,210
- References
- 4
- Citations by year
- 2026: 144 · 2025: 280 · 2024: 313 · 2023: 361 · 2022: 330 · 2021: 319 · 2020: 265 · 2019: 151 · 2018: 47
- FWCI
- 56.51
- Type
- conference paper
- Language
- en
- Open access
- green
- arxiv.org
- Locations
- Interspeech (USB) (publishedVersion)
- arXiv (Cornell University) (submittedVersion)
- arXiv (Cornell University)
- Funders
- Engineering and Physical Sciences Research Council (EP/M013774/1)
- Topics
- Physical Sciences › Computer Science › Artificial Intelligence › Speech Recognition and Synthesis
- Physical Sciences › Computer Science › Signal Processing › Speech and Audio Processing
- Physical Sciences › Computer Science › Computer Vision and Pattern Recognition › Face recognition and analysis
- Keywords
- speaker identification, speaker verification, VoxCeleb, text-independent speaker recognition, automated dataset generation, two-stream CNN, face recognition, YouTube videos
About these checks
Registry, retraction, Cites and statistics results are computed, not judged, with one exception: a printed reference that no index lists is read once by a model before it counts as a citation error. Claims are extracted by a model and kept only when their quote appears verbatim in the text; each cited source is then checked, with a model’s help, for whether it says what the claim relies on. Those verdicts can be wrong — read the claim’s quote and the cited source before relying on one.
A language model drew the claims from the paper on 2026-10-11, under version 3 of the claim definition (model deepseek-v4.1-flash, prompt extract-v3).
A second model labelled each claim on 2026-10-11 (model jev-1.13-20260917, definition revision gold-guideline-v0.2).
Statistics and arithmetic are recomputed, not judged.
Confidence bands are shown only for fields calibrated on our annotated test set. None is calibrated yet, so no band is shown.
What counts as a claim ›How we measure confidence ›How we check this ›
Are you an author, or do you see a problem with this page? Request a review
Data sources
Bibliographic data: Semantic Scholar, OpenAlex, Crossref.
Contains information from the Semantic Scholar Academic Graph Datasets, made available under the ODC Attribution License (ODC-BY 1.0).
Semantic Scholar’s data is described in: Kinney, R., et al. (2023). The Semantic Scholar Open Data Platform. arXiv:2301.10140.