Broad Phonetic Classification of ASR using Visual Based Features
Lehabik, D., Merzban, M., Saad, S., Gody, A. (2020). Broad Phonetic Classification of ASR using Visual Based Features. EKB Journal Management System, 7(1), 14-26. doi: 10.21608/ejle.2020.24358.1003
Doaa Ahmed Lehabik; Mohamed H. Merzban; Sameh F. Saad; Amr M. Gody. "Broad Phonetic Classification of ASR using Visual Based Features". EKB Journal Management System, 7, 1, 2020, 14-26. doi: 10.21608/ejle.2020.24358.1003
Lehabik, D., Merzban, M., Saad, S., Gody, A. (2020). 'Broad Phonetic Classification of ASR using Visual Based Features', EKB Journal Management System, 7(1), pp. 14-26. doi: 10.21608/ejle.2020.24358.1003
Lehabik, D., Merzban, M., Saad, S., Gody, A. Broad Phonetic Classification of ASR using Visual Based Features. EKB Journal Management System, 2020; 7(1): 14-26. doi: 10.21608/ejle.2020.24358.1003

Broad Phonetic Classification of ASR using Visual Based Features

The Egyptian Journal of Language Engineering
Article 2, Volume 7, Issue 1, April 2020, Page 14-26  XML PDF (1.61 MB)
Document Type: Original Article
DOI: 10.21608/ejle.2020.24358.1003
View on SCiNiTO View on SCiNiTO
Authors
Doaa Ahmed Lehabik email 1; Mohamed H. Merzban2; Sameh F. Saad3; Amr M. Godyorcid 4
1Department of Communication and Electronic Faculty of Engineering Fayoum University
2Faculty of Engineering Fayoum University
3Faculty of engineering
4Faculty of Engineering, Fayoum University
Abstract
Abstract: This paper presents a novel method of classifying speech phonemes. Four hybrid techniques based on the acoustic-phonetic approach and pattern recognition approach are used to emphasize the principle idea of this research. The first hybrid model is constructed of fixed state, structured Hidden Markov Model, Gaussian Mixture, Mel scaled Best Tree Image, Convolution Neural network, Vector Quantization (FS-HMM-GM-MBTI-CNN-VQ). The second hybrid model is constructed of variable state, dynamically structured Hidden Markov Model, Gaussian Mixture, Mel scaled Best Tree Image, Convolution Neural network, Vector Quantization (VS-HMM-GM-MBTI-CNN-VQ). The third hybrid model is constructed of fixed state, structured Hidden Markov Model, Gaussian Mixture, Mel scaled Best Tree Image, Convolution Neural network (FS-HMM-GM-MBTI-CNN). The fourth hybrid model is constructed of variable state, dynamically structured Hidden Markov Model, Gaussian Mixture, Mel scaled Best Tree Image, Convolution Neural network (VS-HMM-GM-MBTI-CNN). TIMIT database is used in this paper. All phones are classified into five classes and segregated into Vowels, Plosives, Fricatives, Nasals, and Silences. The results show that using (VS-HMM-GM-MBTI-CNN-VQ) is an available method for classification of phonemes, with the potential for use in applications such as automatic speech recognition and automatic language identification. Competitive results are achieved especially in nasals, plosives, and silence high successive rates than others.
Keywords
ASR; HTK; Convolution Neural Network; Vector Quantization; Hidden Markov Model
Statistics
Article View: 257
PDF Download: 427