Melola Discovery

← Melola Discovery

Jurnal SINTA 2

Improving the Performance of Low-resourced Speaker Identification with Data Preprocessing

Journal of ICT Research and Applications (JICTRA) · Vol. 17 No. 3 (2023) · pp. 275–291

DOI
10.5614/itbj.ict.res.appl.2023.17.3.1
ISSN
2338-5499, 2337-5787
Published
2023-12-31
Publisher
DRI - ITB
Language
eng

Abstract

Automatic speaker identification is done to tackle daily security problems. Speech data collection is an essential but very challenging task for under-resourced languages like Burmese. The speech quality is crucial to accurately recognize the speaker’s identity. This work attempted to find the optimal speech quality appropriate for Burmese tone to enhance identification compared with other more richy resourced languages on Mel-frequency cepstral coefficients (MFCCs). A Burmese speech dataset was created as part of our work because no appropriate dataset available for use. In order to achieve better performance, we preprocessed the foremost recording quality proper for not only Burmese tone but also for nine other Asian languages to achieve multilingual speaker identification. The performance of the preprocessed data was evaluated by comparing with the original data, using a time delay neural network (TDNN) together with a subsampling technique that can reduce time complexity in model training. The experiments were investigated and analyzed on speech datasets of ten Asian languages to reveal the effectiveness of the data preprocessing. The dataset outperformed the original dataset with improvements in terms of equal error rate (EER). The evaluation pointed out that the performance of the system with the preprocessed dataset improved that of the original dataset.

Keywords

Collaboration network of the authors

5 authors · 6 co-authorship links · 2 across different organisations

Hover or tap a name to highlight its co-authors. Double-click to open the author's page.

  • University of Computer Studies Yangon
  • Waseda University
  • Different organisations
  • Same organisation
  • Affiliation unknown
  • Thicker line = more joint articles · bigger circle = more articles
Show as table
Co-authorOrganisationJoint articles
Aye Nyein MonUniversity of Computer Studies Yangon1
Ye Kyaw ThuWaseda University1

Affiliations come from OpenAlex; journal OAI-PMH metadata carries none. Authors without an OpenAlex ID are matched by name, which can merge or split people.

Metadata source: OAI-PMH (metadata jurnal) https://journals.itb.ac.id/index.php/jictra/oai, first harvested 26 Okt 2026, last seen 1 Nov 2026.
Original metadata (oai_dc) 
<oai_dc:dc xmlns="http://www.openarchives.org/OAI/2.0/" xmlns:dc="http://purl.org/dc/elements/1.1/" xmlns:oai_dc="http://www.openarchives.org/OAI/2.0/oai_dc/" xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance" xsi:schemaLocation="http://www.openarchives.org/OAI/2.0/oai_dc/  http://www.openarchives.org/OAI/2.0/oai_dc.xsd">
	<dc:title xml:lang="en-US">Improving the Performance of Low-resourced Speaker Identification with Data Preprocessing</dc:title>
	<dc:creator>Phyu, Win Lai Lai</dc:creator>
	<dc:creator>Naing, Hay Mar Soe</dc:creator>
	<dc:creator>Pa, Win Pa</dc:creator>
	<dc:subject xml:lang="en-US">Burmese speech dataset</dc:subject>
	<dc:subject xml:lang="en-US">data scrutiny</dc:subject>
	<dc:subject xml:lang="en-US">Mel-frequency cepstral coefficients (MFCCs)</dc:subject>
	<dc:subject xml:lang="en-US">multilingual speaker identification</dc:subject>
	<dc:subject xml:lang="en-US">time delay neural network (TDNN)</dc:subject>
	<dc:description xml:lang="en-US">Automatic speaker identification is done to tackle daily security problems. Speech data collection is an essential but very challenging task for under-resourced languages like Burmese. The speech quality is crucial to accurately recognize the speaker’s identity. This work attempted to find the optimal speech quality appropriate for Burmese tone to enhance identification compared with other more richy resourced languages on Mel-frequency cepstral coefficients (MFCCs). A Burmese speech dataset was created as part of our work because no appropriate dataset available for use. In order to achieve better performance, we preprocessed the foremost recording quality proper for not only Burmese tone but also for nine other Asian languages to achieve multilingual speaker identification. The performance of the preprocessed data was evaluated by comparing with the original data, using a time delay neural network (TDNN) together with a subsampling technique that can reduce time complexity in model training. The experiments were investigated and analyzed on speech datasets of ten Asian languages to reveal the effectiveness of the data preprocessing. The dataset outperformed the original dataset with improvements in terms of  equal error rate (EER). The evaluation pointed out that the performance of the system with the preprocessed dataset improved that of the original dataset.</dc:description>
	<dc:publisher xml:lang="en-US">DRI - ITB</dc:publisher>
	<dc:date>2023-12-31</dc:date>
	<dc:type>info:eu-repo/semantics/article</dc:type>
	<dc:type>info:eu-repo/semantics/publishedVersion</dc:type>
	<dc:type xml:lang="en-US">Peer-reviewed Article</dc:type>
	<dc:format>application/pdf</dc:format>
	<dc:identifier>https://journals.itb.ac.id/index.php/jictra/article/view/20340</dc:identifier>
	<dc:identifier>10.5614/itbj.ict.res.appl.2023.17.3.1</dc:identifier>
	<dc:source xml:lang="en-US">Journal of ICT Research and Applications; Vol. 17 No. 3 (2023); 275-291</dc:source>
	<dc:source>2338-5499</dc:source>
	<dc:source>2337-5787</dc:source>
	<dc:source>10.5614/itbj.ict.res.appl.2023.17.3</dc:source>
	<dc:language>eng</dc:language>
	<dc:relation>https://journals.itb.ac.id/index.php/jictra/article/view/20340/6484</dc:relation>
	<dc:rights xml:lang="en-US">Copyright (c) 2023 Journal of ICT Research and Applications</dc:rights>
</oai_dc:dc>