Please use this identifier to cite or link to this item:
https://doi.org/10.21256/zhaw-4254
Publication type: | Conference paper |
Type of review: | Peer review (publication) |
Title: | Speaker clustering using dominant sets |
Authors: | Hibraj, Feliks Vascon, Sebastiano Stadelmann, Thilo Pelillo, Marcello |
DOI: | 10.1109/ICPR.2018.8546067 10.21256/zhaw-4254 |
Proceedings: | 2018 24th International Conference on Pattern Recognition (ICPR) |
Page(s): | 3549 |
Pages to: | 3554 |
Conference details: | 24th International Conference on Pattern Recognition (ICPR 2018), Beijing, China, 20-28 August 2018 |
Issue Date: | 2018 |
Publisher / Ed. Institution: | IEEE |
ISBN: | 978-1-5386-3788-3 |
Language: | English |
Subjects: | Speaker recognition; Speaker embeddings |
Subject (DDC): | 006: Special computer methods |
Abstract: | Speaker clustering is the task of forming speaker-specific groups based on a set of utterances. In this paper, we address this task by using Dominant Sets (DS). DS is a graphbased clustering algorithm with interesting properties that fits well to our problem and has never been applied before to speaker clustering. We report on a comprehensive set of experiments on the TIMIT dataset against standard clustering techniques and specific speaker clustering methods. Moreover, we compare performances under different features by using ones learned via deep neural network directly on TIMIT and other ones extracted from a pre-trained VGGVox net. To asses the stability, we perform a sensitivity analysis on the free parameters of our method, showing that performance is stable under parameter changes. The extensive experimentation carried out confirms the validity of the proposed method, reporting state-of-the-art results under three different standard metrics. We also report reference baseline results for speaker clustering on the entire TIMIT dataset for the first time. |
URI: | https://digitalcollection.zhaw.ch/handle/11475/6081 |
Fulltext version: | Submitted version |
License (according to publishing contract): | Not specified |
Departement: | School of Engineering |
Organisational Unit: | Institute of Computer Science (InIT) |
Appears in collections: | Publikationen School of Engineering |
Files in This Item:
File | Description | Size | Format | |
---|---|---|---|---|
ICPR18b.pdf | 1.1 MB | Adobe PDF | View/Open |
Show full item record
Hibraj, F., Vascon, S., Stadelmann, T., & Pelillo, M. (2018). Speaker clustering using dominant sets [Conference paper]. 2018 24th International Conference on Pattern Recognition (ICPR), 3549–3554. https://doi.org/10.1109/ICPR.2018.8546067
Hibraj, F. et al. (2018) ‘Speaker clustering using dominant sets’, in 2018 24th International Conference on Pattern Recognition (ICPR). IEEE, pp. 3549–3554. Available at: https://doi.org/10.1109/ICPR.2018.8546067.
F. Hibraj, S. Vascon, T. Stadelmann, and M. Pelillo, “Speaker clustering using dominant sets,” in 2018 24th International Conference on Pattern Recognition (ICPR), 2018, pp. 3549–3554. doi: 10.1109/ICPR.2018.8546067.
HIBRAJ, Feliks, Sebastiano VASCON, Thilo STADELMANN und Marcello PELILLO, 2018. Speaker clustering using dominant sets. In: 2018 24th International Conference on Pattern Recognition (ICPR). Conference paper. IEEE. 2018. S. 3549–3554. ISBN 978-1-5386-3788-3
Hibraj, Feliks, Sebastiano Vascon, Thilo Stadelmann, and Marcello Pelillo. 2018. “Speaker Clustering Using Dominant Sets.” Conference paper. In 2018 24th International Conference on Pattern Recognition (ICPR), 3549–54. IEEE. https://doi.org/10.1109/ICPR.2018.8546067.
Hibraj, Feliks, et al. “Speaker Clustering Using Dominant Sets.” 2018 24th International Conference on Pattern Recognition (ICPR), IEEE, 2018, pp. 3549–54, https://doi.org/10.1109/ICPR.2018.8546067.
Items in DSpace are protected by copyright, with all rights reserved, unless otherwise indicated.