The Optimal Performance of Multi-Layer Neural Network for Speaker-Independent Isolated Spoken Malay Parliamentary speech

Seman, Noraini and Abu Bakar, Zainab and Abu Bakar, Nordin and Mohamed, Haslizatul Fairuz and Abdullah, Nur Atiqah Sia and Prasanna, Ramakrisnan and Syed Ahmad, Sharifah Mumtazah (2010) The Optimal Performance of Multi-Layer Neural Network for Speaker-Independent Isolated Spoken Malay Parliamentary speech. Malaysian Journal of Computing (MJoC), 1 (1). pp. 1-9. ISSN 2231-7473

Download

[thumbnail of 11106.pdf] Text
11106.pdf

Download (330kB)

Abstract

This paper describes speech recognizer modeling techniques which are suited to high performance and robust isolated word recognition in speaker-independent manner. In this study, a speech recognition system is presented, specifically for an isolated spoken Malay word recognizer which uses spontaneous and formal speeches collected from Parliament of Malaysia. Currently the vocabulary is limited to ten words that can be pronounced exactly as it written and control the distribution of the vocalic segments. The speech segmentation task is achieved by adopted energy based parameter and zero crossing rate measure with modification to better locates the beginning and ending points of speech from the spoken words. The training and recognition processes are realized by using Multi-layer Perceptron (MLP) Neural Networks with two-layer feedforward network configurations that are trained with stochastic error back-propagation to adjust its weights and biases after presentation of every training data. The Mel-frequency Cepstral Coefficients (MFCCs) has been chosen as speech extraction approach from each segmented utterance as characteristic features for the word recognizer. The MLP performance to determine the optimal cepstral orders and hidden neurons numbers are analyzed. Recognition results showed that the performance of the two-layer network increased as the numbers of hidden neurons increased. Experimental result also showed that the cepstral orders of 12 to 14 were appropriate for the speech feature extraction for the data in this study.

Metadata

Item Type: Article
Creators:
Creators
Email
Seman, Noraini
aini@tmsk.uitm.edu.my
Abu Bakar, Zainab
zainab@tmsk.uitm.edu.my
Abu Bakar, Nordin
nordin@tmsk.uitm.edu.my
Mohamed, Haslizatul Fairuz
fairuz@tmsk.uitm.edu.my
Abdullah, Nur Atiqah Sia
atiqah@tmsk.uitm.edu.my
Prasanna, Ramakrisnan
prasanna@tmsk.uitm.edu.my
Syed Ahmad, Sharifah Mumtazah
smumtazah@uniten.edu.my
Divisions: Universiti Teknologi MARA, Shah Alam > Faculty of Computer and Mathematical Sciences
Journal or Publication Title: Malaysian Journal of Computing (MJoC)
UiTM Journal Collections: UiTM Journal > Malaysian Journal of Computing (MJoC)
ISSN: 2231-7473
Volume: 1
Number: 1
Page Range: pp. 1-9
Official URL: https://mjoc.uitm.edu.my/
Item ID: 11106
Uncontrolled Keywords: Multi-layer Perceptron, Feedforward, Mel-frequency Cepstral Coefficients, Hidden Neuron, Target vector
URI: https://ir.uitm.edu.my/id/eprint/11106

ID Number

11106

Indexing


View in Google Scholar

Edit Item
Edit Item