Deep Learning Based Automatic Speech Recognition for Turkish

Authors : Burak TOMBALOĞLU, Hamit ERDEM

Pages : 725-739

Doi:10.16984/saufenbilder.711888

View : 46 | Download : 9

Publication Date : 2020-08-01

Article Type : Research Paper

Abstract :Using Deep Neural Networks insert ignore into journalissuearticles values(DNN); as an advanced Artificial Neural Networks insert ignore into journalissuearticles values(ANN); has become widespread with the development of computer technology. Although DNN has been applied for solving Automatic Speech Recognition insert ignore into journalissuearticles values(ASR); problem in some languages, DNN-based Turkish Speech Recognition has not been studied extensively. Turkish language is an agglutinative and a phoneme-based language. In this study, a Deep Belief Network insert ignore into journalissuearticles values(DBN); based Turkish phoneme and speech recognizer is developed. The proposed system recognizes words in the system vocabulary and phoneme components of out of vocabulary insert ignore into journalissuearticles values(OOV); words. Sub-word insert ignore into journalissuearticles values(morpheme); based language modelling is implemented into the system. Each phoneme of Turkish language is also modelled as a sub-word in the model. Sub-word insert ignore into journalissuearticles values(morpheme); based language model is widely used for agglutinative languages to prevent excessive vocabulary size. The performance of the suggested DBN based ASR system is compared with the conventional recognition method, GMM insert ignore into journalissuearticles values(Gaussian Mixture Method); based Hidden Markov Model insert ignore into journalissuearticles values(HMM);. Regarding to performance metrics, the recognition rate of Turkish language is improved in compare with previous studies.
Keywords : deep, neural, networks, belief, automatic, speech, recognition, turkish, Deep Belief Networks

ORIGINAL ARTICLE URL