Isolated word recognition using a hybrid neural network

Author

Tabarabaee, V. ; Azimisadjadi, Babak ; Zahirazami, S. Bahram ; Lucas, Caro

Author_Institution

Electron. Res. Center, Sharif Univ. of Technol., Tehran, Iran

Volume

ii

fYear

1994

fDate

19-22 Apr 1994

Abstract

A hybrid neural network is described. It consists of a Kohonen map and a perceptron. The hybrid is proposed firstly for speaker independent, isolated word recognition. However, it may also be used for other classification problems. The novel idea in this system is the usage of a Kohonen map as the feature extractor which converts phonetic similarities of the speech frames into spatial adjacency in the map. This property simplifies the classification task. The system performance was evaluated for recognition of a limited number of Farsi words (numbers “zero” through “ten”). The overall performance of the recognizer showed to be 93.82%

Keywords

feature extraction; natural languages; pattern classification; perceptrons; self-organising feature maps; speech processing; speech recognition; Farsi words; Kohonen map; classification problems; feature extractor; hybrid neural network; isolated word recognition; numbers; perceptron; phonetic similarities; spatial adjacency; speaker independent isolated word recognition; speech frames; system performance; Feature extraction; Isolation technology; Linear predictive coding; Neural networks; Neurons; Oral communication; Speech analysis; Speech recognition; System performance; Unsupervised learning;

fLanguage

English

Publisher

ieee

Conference_Titel

Acoustics, Speech, and Signal Processing, 1994. ICASSP-94., 1994 IEEE International Conference on

Conference_Location

Adelaide, SA

ISSN

1520-6149

Print_ISBN

0-7803-1775-0

Type

conf

DOI

10.1109/ICASSP.1994.389572

Filename

389572