Using Neural Network with Speaker Applications
In Automatic Speech Recognition (ASR) the non-linear data projection provided by a one hidden layer Multilayer Perceptron (MLP), trained to recognize phonemes, and has previous experiments to provide feature enhancement substantially increased ASR performance, especially in noise. Previous att...
Main Author: | |
---|---|
Format: | Article |
Language: | Arabic |
Published: |
College of Science for Women, University of Baghdad
2010-06-01
|
Series: | Baghdad Science Journal |
Subjects: | |
Online Access: | http://bsj.uobaghdad.edu.iq/index.php/BSJ/article/view/1074 |
id |
doaj-588bef38431540d2addc1264fe949843 |
---|---|
record_format |
Article |
spelling |
doaj-588bef38431540d2addc1264fe9498432020-11-25T00:36:58ZaraCollege of Science for Women, University of BaghdadBaghdad Science Journal2078-86652411-79862010-06-017210.21123/bsj.7.2.1076-1081Using Neural Network with Speaker ApplicationsBaghdad Science Journal In Automatic Speech Recognition (ASR) the non-linear data projection provided by a one hidden layer Multilayer Perceptron (MLP), trained to recognize phonemes, and has previous experiments to provide feature enhancement substantially increased ASR performance, especially in noise. Previous attempts to apply an analogous approach to speaker identification have not succeeded in improving performance, except by combining MLP processed features with other features. We present test results for the TIMIT database which show that the advantage of MLP preprocessing for open set speaker identification increases with the number of speakers used to train the MLP and that improved identification is obtained as this number increases beyond sixty. We also present a method for selecting the speakers used for MLP training which further improves identification performance.http://bsj.uobaghdad.edu.iq/index.php/BSJ/article/view/1074"Speaker recognition, data enhancement, MLP" |
collection |
DOAJ |
language |
Arabic |
format |
Article |
sources |
DOAJ |
author |
Baghdad Science Journal |
spellingShingle |
Baghdad Science Journal Using Neural Network with Speaker Applications Baghdad Science Journal "Speaker recognition, data enhancement, MLP" |
author_facet |
Baghdad Science Journal |
author_sort |
Baghdad Science Journal |
title |
Using Neural Network with Speaker Applications |
title_short |
Using Neural Network with Speaker Applications |
title_full |
Using Neural Network with Speaker Applications |
title_fullStr |
Using Neural Network with Speaker Applications |
title_full_unstemmed |
Using Neural Network with Speaker Applications |
title_sort |
using neural network with speaker applications |
publisher |
College of Science for Women, University of Baghdad |
series |
Baghdad Science Journal |
issn |
2078-8665 2411-7986 |
publishDate |
2010-06-01 |
description |
In Automatic Speech Recognition (ASR) the non-linear data projection provided by a one hidden layer Multilayer Perceptron (MLP), trained to recognize phonemes, and has previous experiments to provide feature enhancement substantially increased ASR performance, especially in noise. Previous attempts to apply an analogous approach to speaker identification have not succeeded in improving performance, except by combining MLP processed features with other features. We present test results for the TIMIT database which show that the advantage of MLP preprocessing for open set speaker identification increases with the number of speakers used to train the MLP and that improved identification is obtained as this number increases beyond sixty. We also present a method for selecting the speakers used for MLP training which further improves identification performance. |
topic |
"Speaker recognition, data enhancement, MLP" |
url |
http://bsj.uobaghdad.edu.iq/index.php/BSJ/article/view/1074 |
work_keys_str_mv |
AT baghdadsciencejournal usingneuralnetworkwithspeakerapplications |
_version_ |
1725303350864379904 |