Building a Speech and Text Corpus of Turkish: Large Corpus Collection with Initial Speech Recognition Results
To build automatic speech recognition (ASR) systems with a low word error rate (WER), a large speech and text corpus is needed. Corpus preparation is the first step required for developing an ASR system for a language with few argument speech documents available. Turkish is a language with limited r...
| Main Authors: | , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
MDPI AG
2020-02-01
|
| Series: | Symmetry |
| Subjects: | |
| Online Access: | https://www.mdpi.com/2073-8994/12/2/290 |
| _version_ | 1798003997494214656 |
|---|---|
| author | Huseyin Polat Saadin Oyucu |
| author_facet | Huseyin Polat Saadin Oyucu |
| author_sort | Huseyin Polat |
| collection | DOAJ |
| description | To build automatic speech recognition (ASR) systems with a low word error rate (WER), a large speech and text corpus is needed. Corpus preparation is the first step required for developing an ASR system for a language with few argument speech documents available. Turkish is a language with limited resources for ASR. Therefore, development of a symmetric Turkish transcribed speech corpus according to the high resources languages corpora is crucial for improving and promoting Turkish speech recognition activities. In this study, we constructed a viable alternative to classical transcribed corpus preparation techniques for collecting Turkish speech data. In the presented approach, three different methods were used. In the first step, subtitles, which are mainly supplied for people with hearing difficulties, were used as transcriptions for the speech utterances obtained from movies. In the second step, data were collected via a mobile application. In the third step, a transfer learning approach to the Grand National Assembly of Turkey session records (videotext) was used. We also provide the initial speech recognition results of artificial neural network and Gaussian mixture-model-based acoustic models for Turkish. For training models, the newly collected corpus and other existing corpora published by the Linguistic Data Consortium were used. In light of the test results of the other existing corpora, the current study showed the relative contribution of corpus variability in a symmetric speech recognition task. The decrease in WER after including the new corpus was more evident with increased verified data size, compensating for the status of Turkish as a low resource language. For further studies, the importance of the corpus and language model in the success of the Turkish ASR system is shown. |
| first_indexed | 2024-04-11T12:16:38Z |
| format | Article |
| id | doaj.art-c93f1f67d9244c63b99105be7316b3e5 |
| institution | Directory Open Access Journal |
| issn | 2073-8994 |
| language | English |
| last_indexed | 2024-04-11T12:16:38Z |
| publishDate | 2020-02-01 |
| publisher | MDPI AG |
| record_format | Article |
| series | Symmetry |
| spelling | doaj.art-c93f1f67d9244c63b99105be7316b3e52022-12-22T04:24:18ZengMDPI AGSymmetry2073-89942020-02-0112229010.3390/sym12020290sym12020290Building a Speech and Text Corpus of Turkish: Large Corpus Collection with Initial Speech Recognition ResultsHuseyin Polat0Saadin Oyucu1Department of Computer Engineering, Faculty of Technology, Gazi University, 06560 Ankara, TurkeyDepartment of Computer Engineering, Faculty of Technology, Gazi University, 06560 Ankara, TurkeyTo build automatic speech recognition (ASR) systems with a low word error rate (WER), a large speech and text corpus is needed. Corpus preparation is the first step required for developing an ASR system for a language with few argument speech documents available. Turkish is a language with limited resources for ASR. Therefore, development of a symmetric Turkish transcribed speech corpus according to the high resources languages corpora is crucial for improving and promoting Turkish speech recognition activities. In this study, we constructed a viable alternative to classical transcribed corpus preparation techniques for collecting Turkish speech data. In the presented approach, three different methods were used. In the first step, subtitles, which are mainly supplied for people with hearing difficulties, were used as transcriptions for the speech utterances obtained from movies. In the second step, data were collected via a mobile application. In the third step, a transfer learning approach to the Grand National Assembly of Turkey session records (videotext) was used. We also provide the initial speech recognition results of artificial neural network and Gaussian mixture-model-based acoustic models for Turkish. For training models, the newly collected corpus and other existing corpora published by the Linguistic Data Consortium were used. In light of the test results of the other existing corpora, the current study showed the relative contribution of corpus variability in a symmetric speech recognition task. The decrease in WER after including the new corpus was more evident with increased verified data size, compensating for the status of Turkish as a low resource language. For further studies, the importance of the corpus and language model in the success of the Turkish ASR system is shown.https://www.mdpi.com/2073-8994/12/2/290automatic speech recognitionspeech corpustext corpusdata acquisitionmulti-layer neural networknatural language processing |
| spellingShingle | Huseyin Polat Saadin Oyucu Building a Speech and Text Corpus of Turkish: Large Corpus Collection with Initial Speech Recognition Results Symmetry automatic speech recognition speech corpus text corpus data acquisition multi-layer neural network natural language processing |
| title | Building a Speech and Text Corpus of Turkish: Large Corpus Collection with Initial Speech Recognition Results |
| title_full | Building a Speech and Text Corpus of Turkish: Large Corpus Collection with Initial Speech Recognition Results |
| title_fullStr | Building a Speech and Text Corpus of Turkish: Large Corpus Collection with Initial Speech Recognition Results |
| title_full_unstemmed | Building a Speech and Text Corpus of Turkish: Large Corpus Collection with Initial Speech Recognition Results |
| title_short | Building a Speech and Text Corpus of Turkish: Large Corpus Collection with Initial Speech Recognition Results |
| title_sort | building a speech and text corpus of turkish large corpus collection with initial speech recognition results |
| topic | automatic speech recognition speech corpus text corpus data acquisition multi-layer neural network natural language processing |
| url | https://www.mdpi.com/2073-8994/12/2/290 |
| work_keys_str_mv | AT huseyinpolat buildingaspeechandtextcorpusofturkishlargecorpuscollectionwithinitialspeechrecognitionresults AT saadinoyucu buildingaspeechandtextcorpusofturkishlargecorpuscollectionwithinitialspeechrecognitionresults |