End-to-end neural automatic speech recognition system for low resource languages

IF 5 3区 计算机科学 Q1 COMPUTER SCIENCE, ARTIFICIAL INTELLIGENCE
Sami Dhahbi , Nasir Saleem , Sami Bourouis , Mouhebeddine Berrima , Elena Verdú
{"title":"End-to-end neural automatic speech recognition system for low resource languages","authors":"Sami Dhahbi ,&nbsp;Nasir Saleem ,&nbsp;Sami Bourouis ,&nbsp;Mouhebeddine Berrima ,&nbsp;Elena Verdú","doi":"10.1016/j.eij.2025.100615","DOIUrl":null,"url":null,"abstract":"<div><div>The rising popularity of end-to-end (E2E) automatic speech recognition (ASR) systems can be attributed to their ability to learn complex speech patterns directly from raw data, eliminating the need for intricate feature extraction pipelines and handcrafted language models. E2E-ASR systems have consistently outperformed traditional ASRs. However, training E2E-ASR systems for low-resource languages remains challenging due to the dependence on data from well-resourced languages. ASR is vital for promoting under-resourced languages, especially in developing human-to-human and human-to-machine communication systems. Using synthetic speech and data augmentation techniques can enhance E2E-ASR performance for low-resource languages, reducing word error rates (WERs) and character error rates (CERs). This study leverages a non-autoregressive neural text-to-speech (TTS) engine to generate high-quality speech, converting a series of phonemes into speech waveforms (mel-spectrograms). An on-the-fly data augmentation method is applied to these mel-spectrograms, treating them as images from which features are extracted to train a convolutional neural network (CNN) and a bidirectional long short-term memory (BLSTM)-based ASR. The E2E architecture of this system achieves optimal WER and CER performance. The proposed deep learning-based E2E-ASR, trained with synthetic speech and data augmentation, shows significant performance improvements, with a 20.75% reduction in WERs and a 10.34% reduction in CERs.</div></div>","PeriodicalId":56010,"journal":{"name":"Egyptian Informatics Journal","volume":"29 ","pages":"Article 100615"},"PeriodicalIF":5.0000,"publicationDate":"2025-01-28","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":"0","resultStr":null,"platform":"Semanticscholar","paperid":null,"PeriodicalName":"Egyptian Informatics Journal","FirstCategoryId":"94","ListUrlMain":"https://www.sciencedirect.com/science/article/pii/S1110866525000088","RegionNum":3,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":null,"EPubDate":"","PubModel":"","JCR":"Q1","JCRName":"COMPUTER SCIENCE, ARTIFICIAL INTELLIGENCE","Score":null,"Total":0}
引用次数: 0

Abstract

The rising popularity of end-to-end (E2E) automatic speech recognition (ASR) systems can be attributed to their ability to learn complex speech patterns directly from raw data, eliminating the need for intricate feature extraction pipelines and handcrafted language models. E2E-ASR systems have consistently outperformed traditional ASRs. However, training E2E-ASR systems for low-resource languages remains challenging due to the dependence on data from well-resourced languages. ASR is vital for promoting under-resourced languages, especially in developing human-to-human and human-to-machine communication systems. Using synthetic speech and data augmentation techniques can enhance E2E-ASR performance for low-resource languages, reducing word error rates (WERs) and character error rates (CERs). This study leverages a non-autoregressive neural text-to-speech (TTS) engine to generate high-quality speech, converting a series of phonemes into speech waveforms (mel-spectrograms). An on-the-fly data augmentation method is applied to these mel-spectrograms, treating them as images from which features are extracted to train a convolutional neural network (CNN) and a bidirectional long short-term memory (BLSTM)-based ASR. The E2E architecture of this system achieves optimal WER and CER performance. The proposed deep learning-based E2E-ASR, trained with synthetic speech and data augmentation, shows significant performance improvements, with a 20.75% reduction in WERs and a 10.34% reduction in CERs.
求助全文
约1分钟内获得全文 求助全文
来源期刊
Egyptian Informatics Journal
Egyptian Informatics Journal Decision Sciences-Management Science and Operations Research
CiteScore
11.10
自引率
1.90%
发文量
59
审稿时长
110 days
期刊介绍: The Egyptian Informatics Journal is published by the Faculty of Computers and Artificial Intelligence, Cairo University. This Journal provides a forum for the state-of-the-art research and development in the fields of computing, including computer sciences, information technologies, information systems, operations research and decision support. Innovative and not-previously-published work in subjects covered by the Journal is encouraged to be submitted, whether from academic, research or commercial sources.
×
引用
GB/T 7714-2015
复制
MLA
复制
APA
复制
导出至
BibTeX EndNote RefMan NoteFirst NoteExpress
×
提示
您的信息不完整,为了账户安全,请先补充。
现在去补充
×
提示
您因"违规操作"
具体请查看互助需知
我知道了
×
提示
确定
请完成安全验证×
copy
已复制链接
快去分享给好友吧!
我知道了
右上角分享
点击右上角分享
0
联系我们:info@booksci.cn Book学术提供免费学术资源搜索服务,方便国内外学者检索中英文文献。致力于提供最便捷和优质的服务体验。 Copyright © 2023 布克学术 All rights reserved.
京ICP备2023020795号-1
ghs 京公网安备 11010802042870号
Book学术文献互助
Book学术文献互助群
群 号:481959085
Book学术官方微信