Conference item
VoxCeleb2: Deep speaker recognition
- Abstract:
-
The objective of this paper is speaker recognition under noisy and unconstrained conditions.
We make two key contributions. First, we introduce a very large-scale audio-visual speaker recognition dataset collected from open-source media. Using a fully automated pipeline, we curate VoxCeleb2 which contains over a million utterances from over 6,000 speakers. This is several times larger than any publicly available speaker recognition dataset.
Second, we develop and compare Convolutional Neural Network (CNN) models and training strategies that can effectively recognise identities from voice under various conditions. The models trained on the VoxCeleb2 dataset surpass the performance of previous works on a benchmark dataset by a significant margin.
- Publication status:
- Published
- Peer review status:
- Peer reviewed
Actions
Access Document
- Files:
-
-
(Preview, Accepted manuscript, pdf, 5.4MB, Terms of use)
-
- Publisher copy:
- 10.21437/Interspeech.2018-1929
Authors
- Publisher:
- International Speech Communication Association
- Host title:
- Interspeech 2018
- Journal:
- Interspeech 2018 More from this journal
- Publication date:
- 2018-09-06
- Acceptance date:
- 2018-06-03
- DOI:
- ISSN:
-
1990-9772
- Keywords:
- Pubs id:
-
pubs:944820
- UUID:
-
uuid:08ab75c5-aa1c-49fc-b36a-1280c6a309c4
- Local pid:
-
pubs:944820
- Source identifiers:
-
944820
- Deposit date:
-
2018-11-21
- ARK identifier:
Terms of use
- Copyright holder:
- International Speech Communication Association
- Copyright date:
- 2018
- Notes:
- Copyright © 2018 International Speech Communication Association. This is the accepted manuscript version of the article. The final version is available online from International Speech Communication Association at: http://dx.doi.org/10.21437/Interspeech.2018-1929
If you are the owner of this record, you can report an update to it here: Report update to this record