Hybrid DNN-HMM-Based Approach for Telugu Language Speech Recognition

In this paper, we discuss our research work in Telugu Indian language speech data for building a large language vocabulary to build Telugu speech recognition system. We have collected speech research data around 2628 utterances in Telugu language among 14 Indian regional languages as part of Speech...

Full description

Saved in:
Bibliographic Details
Main Authors: M. Rama Rajeswari, Suryakanth V. Gangashetty
Format: Article
Language:English
Published: IEEE 2025-01-01
Series:IEEE Access
Subjects:
Online Access:https://ieeexplore.ieee.org/document/11079596/
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1849714630247579648
author M. Rama Rajeswari
Suryakanth V. Gangashetty
author_facet M. Rama Rajeswari
Suryakanth V. Gangashetty
author_sort M. Rama Rajeswari
collection DOAJ
description In this paper, we discuss our research work in Telugu Indian language speech data for building a large language vocabulary to build Telugu speech recognition system. We have collected speech research data around 2628 utterances in Telugu language among 14 Indian regional languages as part of Speech Quality Control Project under Bhashini-NLTM (National Language Translation Mission) in collaboration with IIT Madras. In this research work, we discuss our methodology for the collection of Telugu data and developing Telugu speech database. In our paper, we present the preliminary results that are obtained by building speech recognition using the acoustic models created on Telugu language database by implementing Kaldi toolkit for speech in building ASR in Telugu language. The proposed hybrid DNN-HMM acoustic model implemented to build Telugu ASR in Kaldi achieved a WER score of 0. 7% when the weight of the inverse_language_model is 17 in the test data set of the Telugu speech database. In the process of our work, some of the ASR models, methods and speech processing performance metrics that we used in ASR in order to maintain quality corpus are discussed.
format Article
id doaj-art-2daaf9e5d5354b95896da164d4dcbfe7
institution DOAJ
issn 2169-3536
language English
publishDate 2025-01-01
publisher IEEE
record_format Article
series IEEE Access
spelling doaj-art-2daaf9e5d5354b95896da164d4dcbfe72025-08-20T03:13:39ZengIEEEIEEE Access2169-35362025-01-011312275212276810.1109/ACCESS.2025.358866411079596Hybrid DNN-HMM-Based Approach for Telugu Language Speech RecognitionM. Rama Rajeswari0https://orcid.org/0009-0002-8997-5527Suryakanth V. Gangashetty1https://orcid.org/0000-0001-6745-4363Department of Computer Science and Engineering, Koneru Lakshmaiah Education Foundation, Guntur, Andhra Pradesh, IndiaDepartment of Computer Science and Engineering, Koneru Lakshmaiah Education Foundation, Guntur, Andhra Pradesh, IndiaIn this paper, we discuss our research work in Telugu Indian language speech data for building a large language vocabulary to build Telugu speech recognition system. We have collected speech research data around 2628 utterances in Telugu language among 14 Indian regional languages as part of Speech Quality Control Project under Bhashini-NLTM (National Language Translation Mission) in collaboration with IIT Madras. In this research work, we discuss our methodology for the collection of Telugu data and developing Telugu speech database. In our paper, we present the preliminary results that are obtained by building speech recognition using the acoustic models created on Telugu language database by implementing Kaldi toolkit for speech in building ASR in Telugu language. The proposed hybrid DNN-HMM acoustic model implemented to build Telugu ASR in Kaldi achieved a WER score of 0. 7% when the weight of the inverse_language_model is 17 in the test data set of the Telugu speech database. In the process of our work, some of the ASR models, methods and speech processing performance metrics that we used in ASR in order to maintain quality corpus are discussed.https://ieeexplore.ieee.org/document/11079596/ASRHMMMFCCTelugu languageacoustic modellanguage model
spellingShingle M. Rama Rajeswari
Suryakanth V. Gangashetty
Hybrid DNN-HMM-Based Approach for Telugu Language Speech Recognition
IEEE Access
ASR
HMM
MFCC
Telugu language
acoustic model
language model
title Hybrid DNN-HMM-Based Approach for Telugu Language Speech Recognition
title_full Hybrid DNN-HMM-Based Approach for Telugu Language Speech Recognition
title_fullStr Hybrid DNN-HMM-Based Approach for Telugu Language Speech Recognition
title_full_unstemmed Hybrid DNN-HMM-Based Approach for Telugu Language Speech Recognition
title_short Hybrid DNN-HMM-Based Approach for Telugu Language Speech Recognition
title_sort hybrid dnn hmm based approach for telugu language speech recognition
topic ASR
HMM
MFCC
Telugu language
acoustic model
language model
url https://ieeexplore.ieee.org/document/11079596/
work_keys_str_mv AT mramarajeswari hybriddnnhmmbasedapproachfortelugulanguagespeechrecognition
AT suryakanthvgangashetty hybriddnnhmmbasedapproachfortelugulanguagespeechrecognition