Wavelet transform based features vector extraction in isolated words speech recognition system

Author

Al-Qaraawi, Salih M. ; Mahmood, Sarah Shukur

Author_Institution

Comput. Eng. Dept., Univ. of Technol., Baghdad, Iraq

fYear

2014

fDate

23-25 July 2014

Firstpage

847

Lastpage

850

Abstract

The aim of this paper is to develop an algorithm that uses wavelet transform and energy to extract and represent features of the acquired speech signals as a basis for accurate method of identifying and classifying speech signals according to their features. Feed-forward neural network is used in recognition of English spoken words with back propagation learning algorithm. It is important to mention that due to its generality, this method can be applied to recognise many languages.

Keywords

backpropagation; feature extraction; feedforward neural nets; signal classification; signal detection; speech recognition; vectors; wavelet transforms; English spoken word recognition; backpropagation learning algorithm; energy extraction; feature vector extraction; feedforward neural network; isolated word speech recognition system; language recognition; speech signal acquisition; speech signal classification; speech signal identification; wavelet transform; Acoustics; Discrete wavelet transforms; Feature extraction; Speech; Speech recognition; acoustic feature; neural network; speech recognition; wavelet transform;

fLanguage

English

Publisher

ieee

Conference_Titel

Communication Systems, Networks & Digital Signal Processing (CSNDSP), 2014 9th International Symposium on

Conference_Location

Manchester

Type

conf

DOI

10.1109/CSNDSP.2014.6923945

Filename

6923945