NN speech recognition utilizing aligned DTW local distance scores

This paper presents the neural network (NN) speech recognition using processed LPC input features. But NN has a limitation that the network must have a fixed amount of input nodes. The input feature processing method will use frame matching based on Dynamic Time Warping (DTW) algorithm to fix the in...

Full description

Bibliographic Details
Main Authors: Sudirman, Rubita, Salleh, Sh-Hussain, Ting, Chee Ming
Format: Conference or Workshop Item
Language:English
Published: Universiti Teknologi Malaysia 2005
Subjects:
Online Access:http://eprints.utm.my/1699/
http://eprints.utm.my/1699/1/rubita05_ICMT-192.pdf
_version_ 1848890191722315776
author Sudirman, Rubita
Salleh, Sh-Hussain
Ting, Chee Ming
author_facet Sudirman, Rubita
Salleh, Sh-Hussain
Ting, Chee Ming
author_sort Sudirman, Rubita
building UTeM Institutional Repository
collection Online Access
description This paper presents the neural network (NN) speech recognition using processed LPC input features. But NN has a limitation that the network must have a fixed amount of input nodes. The input feature processing method will use frame matching based on Dynamic Time Warping (DTW) algorithm to fix the input size to a fix amount of input vectors. The LPC features are aligned between the input frames (test set) to the reference (training set) using our DTW fixing frame (DTW-FF) algorithm. This proper time normalization is needed since NN is designed to compare data of the same length, whilst same speech can varies in their length. By doing frame fixing or also known as time normalization, the test set and the training set frames are adjusted so that both sets will have the same number of frames according to the reference set. The neural network with backpropagation algorithm is used as the recognition engine at the back-end processing to enhance the recognition performance. The results compare DTW with LPC coefficients to back-propagation NN with LPC coefficients adjusted using DTW.
first_indexed 2025-11-15T20:38:09Z
format Conference or Workshop Item
id utm-1699
institution Universiti Teknologi Malaysia
institution_category Local University
language English
last_indexed 2025-11-15T20:38:09Z
publishDate 2005
publisher Universiti Teknologi Malaysia
recordtype eprints
repository_type Digital Repository
spelling utm-16992017-08-28T00:14:48Z http://eprints.utm.my/1699/ NN speech recognition utilizing aligned DTW local distance scores Sudirman, Rubita Salleh, Sh-Hussain Ting, Chee Ming TK Electrical engineering. Electronics Nuclear engineering This paper presents the neural network (NN) speech recognition using processed LPC input features. But NN has a limitation that the network must have a fixed amount of input nodes. The input feature processing method will use frame matching based on Dynamic Time Warping (DTW) algorithm to fix the input size to a fix amount of input vectors. The LPC features are aligned between the input frames (test set) to the reference (training set) using our DTW fixing frame (DTW-FF) algorithm. This proper time normalization is needed since NN is designed to compare data of the same length, whilst same speech can varies in their length. By doing frame fixing or also known as time normalization, the test set and the training set frames are adjusted so that both sets will have the same number of frames according to the reference set. The neural network with backpropagation algorithm is used as the recognition engine at the back-end processing to enhance the recognition performance. The results compare DTW with LPC coefficients to back-propagation NN with LPC coefficients adjusted using DTW. Universiti Teknologi Malaysia 2005-12-04 Conference or Workshop Item PeerReviewed application/pdf en http://eprints.utm.my/1699/1/rubita05_ICMT-192.pdf Sudirman, Rubita and Salleh, Sh-Hussain and Ting, Chee Ming (2005) NN speech recognition utilizing aligned DTW local distance scores. In: Proceeding of the 9th International Conference on Mechatronics Technology, 5-8 December 2005, Kuala Lumpur. https://books.google.com.my/books/about/ICMT_2005.html?id=qj0UDAEACAAJ&redir_esc=y
spellingShingle TK Electrical engineering. Electronics Nuclear engineering
Sudirman, Rubita
Salleh, Sh-Hussain
Ting, Chee Ming
NN speech recognition utilizing aligned DTW local distance scores
title NN speech recognition utilizing aligned DTW local distance scores
title_full NN speech recognition utilizing aligned DTW local distance scores
title_fullStr NN speech recognition utilizing aligned DTW local distance scores
title_full_unstemmed NN speech recognition utilizing aligned DTW local distance scores
title_short NN speech recognition utilizing aligned DTW local distance scores
title_sort nn speech recognition utilizing aligned dtw local distance scores
topic TK Electrical engineering. Electronics Nuclear engineering
url http://eprints.utm.my/1699/
http://eprints.utm.my/1699/
http://eprints.utm.my/1699/1/rubita05_ICMT-192.pdf