Please use this identifier to cite or link to this item: https://doi.org/10.21256/zhaw-3762
Publication type: Conference paper
Type of review: Peer review (publication)
Title: Learning embeddings for speaker clustering based on voice equality
Authors: Lukic, Yanick X.
Vogt, Carlo
Dürr, Oliver
Stadelmann, Thilo
DOI: 10.21256/zhaw-3762
10.1109/MLSP.2017.8168166
Proceedings: 2017 IEEE 27th International Workshop on Machine Learning for Signal Processing (MLSP)
Conference details: 27th IEEE International Workshop on Machine Learning for Signal Processing (MLSP 2017), Tokyo, 25-28 September 2017
Issue Date: 2017
Publisher / Ed. Institution: IEEE
ISBN: 978-1-5090-6341-3
Other identifiers: INSPEC Accession Number: 17416144
Language: English
Subjects: Datalab; Speaker recognition; Deep learning; Speaker clustering
Subject (DDC): 006: Special computer methods
Abstract: Recent work has shown that convolutional neural networks (CNNs) trained in a supervised fashion for speaker identification are able to extract features from spectrograms which can be used for speaker clustering. These features are represented by the activations of a certain hidden layer and are called embeddings. However, previous approaches require plenty of additional speaker data to learn the embedding, and although the clustering results are then on par with more traditional approaches using MFCC features etc., room for improvements stems from the fact that these embeddings are trained with a surrogate task that is rather far away from segregating unknown voices - namely, identifying few specific speakers. We address both problems by training a CNN to extract embeddings that are similar for equal speakers (regardless of their specific identity) using weakly labeled data. We demonstrate our approach on the well-known TIMIT dataset that has often been used for speaker clustering experiments in the past. We exceed the clustering performance of all previous approaches, but require just 100 instead of 590 unrelated speakers to learn an embedding suited for clustering.
URI: https://digitalcollection.zhaw.ch/handle/11475/7088
Fulltext version: Accepted version
License (according to publishing contract): Licence according to publishing contract
Departement: School of Engineering
Organisational Unit: Institute of Computer Science (InIT)
Institute of Data Analysis and Process Design (IDP)
Appears in collections:Publikationen School of Engineering

Files in This Item:
File Description SizeFormat 
MLSP_2017.pdf1.37 MBAdobe PDFThumbnail
View/Open
Show full item record
Lukic, Y. X., Vogt, C., Dürr, O., & Stadelmann, T. (2017). Learning embeddings for speaker clustering based on voice equality. 2017 IEEE 27th International Workshop on Machine Learning for Signal Processing (MLSP). https://doi.org/10.21256/zhaw-3762
Lukic, Y.X. et al. (2017) ‘Learning embeddings for speaker clustering based on voice equality’, in 2017 IEEE 27th International Workshop on Machine Learning for Signal Processing (MLSP). IEEE. Available at: https://doi.org/10.21256/zhaw-3762.
Y. X. Lukic, C. Vogt, O. Dürr, and T. Stadelmann, “Learning embeddings for speaker clustering based on voice equality,” in 2017 IEEE 27th International Workshop on Machine Learning for Signal Processing (MLSP), 2017. doi: 10.21256/zhaw-3762.
LUKIC, Yanick X., Carlo VOGT, Oliver DÜRR und Thilo STADELMANN, 2017. Learning embeddings for speaker clustering based on voice equality. In: 2017 IEEE 27th International Workshop on Machine Learning for Signal Processing (MLSP). Conference paper. IEEE. 2017. ISBN 978-1-5090-6341-3
Lukic, Yanick X., Carlo Vogt, Oliver Dürr, and Thilo Stadelmann. 2017. “Learning Embeddings for Speaker Clustering Based on Voice Equality.” Conference paper. In 2017 IEEE 27th International Workshop on Machine Learning for Signal Processing (MLSP). IEEE. https://doi.org/10.21256/zhaw-3762.
Lukic, Yanick X., et al. “Learning Embeddings for Speaker Clustering Based on Voice Equality.” 2017 IEEE 27th International Workshop on Machine Learning for Signal Processing (MLSP), IEEE, 2017, https://doi.org/10.21256/zhaw-3762.


Items in DSpace are protected by copyright, with all rights reserved, unless otherwise indicated.