Building a Speech and Text Corpus of Turkish: Large Corpus Collection with Initial Speech Recognition Results

Abstract

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.

Braille Recognition for Reducing Asymmetric Communication between the Blind and Non-Blind

“…The dataset consists of English braille symbols, specifically English Grade 1 braille which is un-contracted, meaning there is a specific braille symbol for each letter in the English alphabet as well as for punctuation. Just as there are insufficient datasets for speech-to-text recognition [25], multiple online searches showed a lack of braille character datasets that fit the needs of this study. The few that do exist are either in foreign languages, have broken links, are not publicly available, or only contain printed images of braille, which are not suitable for training a recognition system that can learn imperfect and blurry images.…”

Türkçe Otomatik Konuşma Tanıma Sistemi için Dil Modeli Optimizasyon Yöntemi

“…The Turkish speech data set [51] prepared by Bogazici University in 2012 and presented by LDC, the Linguistic Data Consortium, and METU 1.0 sound corpus provided by METU were used for training and testing processes of the ASR system [52]. Also, a new corpus was used to exhibit the performance of the method we propose more clearly (HS Corpus) [53]. The corpus information is given in Table 1.…”

Resources for Turkish Natural Language Processing: A critical survey

“…Arısoy et al (2009) report a larger dataset of broadcast news, and a dataset of 38 000 hours of call center recordings is reported by Haznedaroğlu and Arslan (2014). A recent speech corpus, consisting of movies with aligned subtitles, and read speech samples are reported by Polat and Oyucu (2020). The availability of corpora listed above is unclear.…”