Parametric feature-based voice recognition system using artificial neural network

M. Bodrzzaman; K. Kuah; T. Jamil; C. Wang; X. Li

Parametric feature-based voice recognition system using artificial neural network

M. Bodrzzaman^*, K. Kuah, T. Jamil, C. Wang, X. Li

^*Corresponding author for this work

Electrical and Computer Engineering

Research output: Chapter in Book/Report/Conference proceeding › Conference contribution

1 Citation (Scopus)

Abstract

In this paper, a speaker identification experiment was conducted using an artificial neural network. The speech data were collected from nine different speakers saying the same word ″Hello″. The speech data were then preprocessed for signal conditioning. A total of 14 feature parameters were obtained in which twelve of them are the coefficient of the 12^th order linear predictor (LPC) and the other two were selected as the peak and bandwidth of spectral envelop. These 14 feature parameters then served as the input to the neural network for speaker classification. A standard two-layer feedforward neural network was trained to identify different feature sets associated with the corresponding speakers. The network size was selected to be 14-8-4 (14 input, 8 hidden and 4 output units). Nine (9) utterances from each speaker were used as training data and the other one served as testing data. No pagination in original publication.

Original language	English
Title of host publication	Conference Proceedings - IEEE SOUTHEASTCON
Publisher	Publ by IEEE
ISBN (Print)	0780312570
Publication status	Published - 1993
Event	Proceedings of the IEEE Southeastcon '93 - Charlotte, NC, USA Duration: Apr 4 1993 → Apr 7 1993

Other

Other	Proceedings of the IEEE Southeastcon '93
City	Charlotte, NC, USA
Period	4/4/93 → 4/7/93

ASJC Scopus subject areas

Electrical and Electronic Engineering

Cite this

@inproceedings{031141c9d3234045b633c2a46ad155e4,

title = "Parametric feature-based voice recognition system using artificial neural network",

abstract = "In this paper, a speaker identification experiment was conducted using an artificial neural network. The speech data were collected from nine different speakers saying the same word ″Hello″. The speech data were then preprocessed for signal conditioning. A total of 14 feature parameters were obtained in which twelve of them are the coefficient of the 12th order linear predictor (LPC) and the other two were selected as the peak and bandwidth of spectral envelop. These 14 feature parameters then served as the input to the neural network for speaker classification. A standard two-layer feedforward neural network was trained to identify different feature sets associated with the corresponding speakers. The network size was selected to be 14-8-4 (14 input, 8 hidden and 4 output units). Nine (9) utterances from each speaker were used as training data and the other one served as testing data. No pagination in original publication.",

author = "M. Bodrzzaman and K. Kuah and T. Jamil and C. Wang and X. Li",

year = "1993",

language = "English",

isbn = "0780312570",

booktitle = "Conference Proceedings - IEEE SOUTHEASTCON",

publisher = "Publ by IEEE",

note = "Proceedings of the IEEE Southeastcon '93 ; Conference date: 04-04-1993 Through 07-04-1993",

}

TY - GEN

T1 - Parametric feature-based voice recognition system using artificial neural network

AU - Bodrzzaman, M.

AU - Kuah, K.

AU - Jamil, T.

AU - Wang, C.

AU - Li, X.

PY - 1993

Y1 - 1993

N2 - In this paper, a speaker identification experiment was conducted using an artificial neural network. The speech data were collected from nine different speakers saying the same word ″Hello″. The speech data were then preprocessed for signal conditioning. A total of 14 feature parameters were obtained in which twelve of them are the coefficient of the 12th order linear predictor (LPC) and the other two were selected as the peak and bandwidth of spectral envelop. These 14 feature parameters then served as the input to the neural network for speaker classification. A standard two-layer feedforward neural network was trained to identify different feature sets associated with the corresponding speakers. The network size was selected to be 14-8-4 (14 input, 8 hidden and 4 output units). Nine (9) utterances from each speaker were used as training data and the other one served as testing data. No pagination in original publication.

AB - In this paper, a speaker identification experiment was conducted using an artificial neural network. The speech data were collected from nine different speakers saying the same word ″Hello″. The speech data were then preprocessed for signal conditioning. A total of 14 feature parameters were obtained in which twelve of them are the coefficient of the 12th order linear predictor (LPC) and the other two were selected as the peak and bandwidth of spectral envelop. These 14 feature parameters then served as the input to the neural network for speaker classification. A standard two-layer feedforward neural network was trained to identify different feature sets associated with the corresponding speakers. The network size was selected to be 14-8-4 (14 input, 8 hidden and 4 output units). Nine (9) utterances from each speaker were used as training data and the other one served as testing data. No pagination in original publication.

UR - http://www.scopus.com/inward/record.url?scp=0027151282&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=0027151282&partnerID=8YFLogxK

M3 - Conference contribution

AN - SCOPUS:0027151282

SN - 0780312570

BT - Conference Proceedings - IEEE SOUTHEASTCON

PB - Publ by IEEE

T2 - Proceedings of the IEEE Southeastcon '93

Y2 - 4 April 1993 through 7 April 1993

ER -

Parametric feature-based voice recognition system using artificial neural network

Abstract

Other

ASJC Scopus subject areas

Other files and links

Fingerprint

Cite this