Filters








29 Hits in 2.3 sec

VoxSRC 2020: The Second VoxCeleb Speaker Recognition Challenge [article]

Arsha Nagrani, Joon Son Chung, Jaesung Huh, Andrew Brown, Ernesto Coto, Weidi Xie, Mitchell McLaren, Douglas A Reynolds, Andrew Zisserman
2020 arXiv   pre-print
We held the second installment of the VoxCeleb Speaker Recognition Challenge in conjunction with Interspeech 2020.  ...  The goal of this challenge was to assess how well current speaker recognition technology is able to diarise and recognize speakers in unconstrained or 'in the wild' data.  ...  INTRODUCTION In 2019 we introduced the VoxCeleb Speaker Recognition Challenge [1] (VoxSRC), a new series of speaker recognition challenges that are intended to be hosted annually.  ... 
arXiv:2012.06867v1 fatcat:k4jmpz7bsncwjdmlzql7vbhmge

The UPC Speaker Verification System Submitted to VoxCeleb Speaker Recognition Challenge 2020 (VoxSRC-20) [article]

Umair Khan, Javier Hernando
2020 arXiv   pre-print
This report describes the submission from Technical University of Catalonia (UPC) to the VoxCeleb Speaker Recognition Challenge (VoxSRC-20) at Interspeech 2020.  ...  Whereas, our triple-branch siamese is trained to learn speaker embeddings using triplet loss. We provide results of our systems on VoxCeleb-1 test, VoxSRC-20 validation and test sets.  ...  The VoxSRC-20 is the second edition of the speaker recognition challenge held by VoxCeleb team.  ... 
arXiv:2010.10937v2 fatcat:mlh6xb3udjhdne646tbhtjj46u

The ins and outs of speaker recognition: lessons from VoxSRC 2020 [article]

Yoohwan Kwon, Hee-Soo Heo, Bong-Jin Lee, Joon Son Chung
2020 arXiv   pre-print
The VoxCeleb Speaker Recognition Challenge (VoxSRC) at Interspeech 2020 offers a challenging evaluation for speaker recognition systems, which includes celebrities playing different parts in movies.  ...  The goal of this work is robust speaker recognition of utterances recorded in these challenging environments.  ...  We would like to thank Brecht Desplanques for his help with the implementation of ECAPA-TDNN.  ... 
arXiv:2010.15809v1 fatcat:uhxg2ni46vbvjbkvulz23k2kca

VoxSRC 2021: The Third VoxCeleb Speaker Recognition Challenge [article]

Andrew Brown, Jaesung Huh, Joon Son Chung, Arsha Nagrani, Andrew Zisserman
2022 arXiv   pre-print
The third instalment of the VoxCeleb Speaker Recognition Challenge was held in conjunction with Interspeech 2021.  ...  The aim of this challenge was to assess how well current speaker recognition technology is able to diarise and recognise speakers in unconstrained or 'in the wild' data.  ...  Acknowledgements This work is funded by the EPSRC programme grant EP/T028572/1 VisualAI. Andrew Brown is funded by an EP-SRC DTA Studentship. Jaesung Huh is funded by a Global Korea Scholarship.  ... 
arXiv:2201.04583v1 fatcat:yyrxdx3yszdd7nosgubhb3lr6m

Clova Baseline System for the VoxCeleb Speaker Recognition Challenge 2020 [article]

Hee Soo Heo, Bong-Jin Lee, Jaesung Huh, Joon Son Chung
2020 arXiv   pre-print
This report describes our submission to the VoxCeleb Speaker Recognition Challenge (VoxSRC) at Interspeech 2020.  ...  We perform a careful analysis of speaker recognition models based on the popular ResNet architecture, and train a number of variants using a range of loss functions.  ...  Introduction The VoxCeleb Speaker Recognition Challenge 2020 is second installment of the new series of speaker recognition challenges that are hosted annually.  ... 
arXiv:2009.14153v1 fatcat:iwvxyaolq5ffbbd7hwvfxjm6ey

The IDLAB VoxSRC-20 Submission: Large Margin Fine-Tuning and Quality-Aware Score Calibration in DNN Based Speaker Verification [article]

Jenthe Thienpondt, Brecht Desplanques, Kris Demuynck
2021 arXiv   pre-print
tracks of the VoxCeleb Speaker Recognition Challenge 2020.  ...  It enables the network to create more robust speaker embeddings by enabling the use of longer training utterances in combination with a more aggressive margin penalty.  ...  For the VoxSRC-20 [30] test set results reported in Table 3 , the MinDCF is evaluated as defined in the challenge with a Ptarget value of 0.05.  ... 
arXiv:2010.11255v2 fatcat:3n5jnj7c7ra23lgjdgs4h55aoq

The xx205 System for the VoxCeleb Speaker Recognition Challenge 2020 [article]

Xu Xiang
2020 arXiv   pre-print
This report describes the systems submitted to the first and second tracks of the VoxCeleb Speaker Recognition Challenge (VoxSRC) 2020, which ranked second in both tracks.  ...  angular margin softmax loss to train the speaker models, and (3) applying score normalization and system fusion to boost the performance.  ...  Introduction The first and second tracks of the VoxSRC 2020 challenge allow participants to train the speaker model in a supervised manner.  ... 
arXiv:2011.00200v1 fatcat:pqj72uhfc5echmx53hk7xwscu4

Tongji University Team for the VoxCeleb Speaker Recognition Challenge 2020 [article]

Rui Wang, Zhihua Wei, Yibin Zhan, Zhuoxi Chen
2020 arXiv   pre-print
In this report, we describe the submission of Tongji University team to the CLOSE track of the VoxCeleb Speaker Recognition Challenge (VoxSRC) 2020 at Interspeech 2020.  ...  We investigate different speaker recognition systems based on the popular ResNet-34 architecture, and train multiple variants via various loss functions.  ...  Introduction The VoxCeleb Speaker Recognition Challenge (VoxSRC) 2020 is second installment of the new series of speaker recognition challenges that are hosted annually.  ... 
arXiv:2010.08179v1 fatcat:j4u7th4s3vd7rngu2qyo2yfy34

ShaneRun System Description to VoxCeleb Speaker Recognition Challenge 2020 [article]

Shen Chen
2020 arXiv   pre-print
In this report, we describe the submission of ShaneRun's team to the VoxCeleb Speaker Recognition Challenge (VoxSRC) 2020.  ...  We use ResNet-34 as encoder to extract the speaker embeddings, which is referenced from the open-source voxceleb-trainer.  ...  Introduction The VoxCeleb Speaker Recognition Challenge 2020 is second term of the new series of speaker recognition challenges.  ... 
arXiv:2011.01518v1 fatcat:tkos7xxuozcafidopckxzis5ka

Microsoft Speaker Diarization System for the VoxCeleb Speaker Recognition Challenge 2020 [article]

Xiong Xiao, Naoyuki Kanda, Zhuo Chen, Tianyan Zhou, Takuya Yoshioka, Sanyuan Chen, Yong Zhao, Gang Liu, Yu Wu, Jian Wu, Shujie Liu, Jinyu Li (+1 others)
2020 arXiv   pre-print
This paper describes the Microsoft speaker diarization system for monaural multi-talker recordings in the wild, evaluated at the diarization track of the VoxCeleb Speaker Recognition Challenge(VoxSRC)  ...  2020.  ...  1st at the VoxSRC challenge 2020.  ... 
arXiv:2010.11458v2 fatcat:lstlnx5udjfbjlijufufsesfky

Tongji University Undergraduate Team for the VoxCeleb Speaker Recognition Challenge2020 [article]

Shufan Shen, Ran Miao, Yi Wang, Zhihua Wei
2020 arXiv   pre-print
In this report, we discribe the submission of Tongji University undergraduate team to the CLOSE track of the VoxCeleb Speaker Recognition Challenge (VoxSRC) 2020 at Interspeech 2020.  ...  Our fusion of two selected systems for the CLOSE track achieves 0.2973 DCF and 4.9700\% EER on the challenge evaluation set.  ...  The authors would like to thank the organizing committees of the INTERSPEECH conferences for providing participant with the template files and the Naver Corporation for giving training framework and pre-trained  ... 
arXiv:2010.10145v1 fatcat:3ij6anxs3rhdfosp3bgbckfvre

The Idlab Voxsrc-20 Submission: Large Margin Fine-Tuning and Quality-Aware Score Calibration in DNN Based Speaker Verification

Jenthe Thienpondt, Brecht Desplanques, Kris Demuynck
2021 ICASSP 2021 - 2021 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP)  
tracks of the Vox-Celeb Speaker Recognition Challenge 2020.  ...  It enables the network to create more robust speaker embeddings by enabling the use of longer training utterances in combination with a more aggressive margin penalty.  ...  For the VoxSRC-20 [30] test set results reported in Table 3 , the MinDCF is evaluated as defined in the challenge with a Ptarget value of 0.05.  ... 
doi:10.1109/icassp39728.2021.9414600 fatcat:zil3jqjohnh7zojziimsujdqy4

The DKU-DukeECE Systems for VoxCeleb Speaker Recognition Challenge 2020 [article]

Weiqing Wang, Danwei Cai, Xiaoyi Qin, Ming Li
2020 arXiv   pre-print
In this paper, we present the system submission for the VoxCeleb Speaker Recognition Challenge 2020 (VoxSRC-20) by the DKU-DukeECE team.  ...  For track 4, we investigate the whole system pipeline for speaker diarization, including voice activity detection (VAD), uniform segmentation, speaker embedding extraction, and clustering.  ...  Moreover, we adopt the speed perturbation using sox to increase the speaker number. The strategy also has a successful application in speech and speaker recognition tasks [5, 6] .  ... 
arXiv:2010.12731v1 fatcat:pleqx3wewbcozdbbobzwrxkfvy

ECAPA-TDNN: Emphasized Channel Attention, Propagation and Aggregation in TDNN Based Speaker Verification

Brecht Desplanques, Jenthe Thienpondt, Kris Demuynck
2020 Interspeech 2020  
The proposed ECAPA-TDNN architecture significantly outperforms state-ofthe-art TDNN based systems on the VoxCeleb test sets and the 2019 VoxCeleb Speaker Recognition Challenge.  ...  Current speaker verification techniques rely on a neural network to extract speaker representations.  ...  However, experiments during the recently held Short-duration Speaker Verification (SdSV) Challenge 2020 [28] convinced us to incorporate summed residuals in the final ECAPA-TDNN architecture.  ... 
doi:10.21437/interspeech.2020-2650 dblp:conf/interspeech/DesplanquesTD20 fatcat:os4wy7dwljgslm6bicxh5px4vi

ECAPA-TDNN: Emphasized Channel Attention, Propagation and Aggregation in TDNN Based Speaker Verification [article]

Brecht Desplanques, Jenthe Thienpondt, Kris Demuynck
2020 arXiv   pre-print
The proposed ECAPA-TDNN architecture significantly outperforms state-of-the-art TDNN based systems on the VoxCeleb test sets and the 2019 VoxCeleb Speaker Recognition Challenge.  ...  Current speaker verification techniques rely on a neural network to extract speaker representations.  ...  However, experiments during the recently held Short-duration Speaker Verification (SdSV) Challenge 2020 [28] convinced us to incorporate summed residuals in the final ECAPA-TDNN architecture.  ... 
arXiv:2005.07143v3 fatcat:jktx5lialjasjkk3t4k3iw5y5e
« Previous Showing results 1 — 15 out of 29 results