|2/2006 - 7|
Noisy Voice Pattern Recognition by Three-Rate Hybrid Neural NetworkIgor ASTROV, Svetlana TATARLY, Sergei TATARLY
|Click to see author's profile on SCOPUS, IEEE Xplore, Web of Science|
|Not available online | Views: 707|
Simulation, multi-rate, neural networks, voice pattern, pattern recognition
References keywords will be displayed on the next page reload.
About this article
Date of Publication: 2006-10-02
Volume 6, Issue 2, Year 2006, On page(s): 35 - 38
ISSN: 1582-7445, e-ISSN: 1844-7600
Digital Object Identifier: Not assigned
This paper presents the three-rate hybrid neural network (TRHNN) for processing of noisy signal. The received TRHNN consists of /fast/ back-propagation neural network (FBPNN), /intermediate/ back-propagation neural network (IBPNN) and /slow/ Kohonen neural network (SKNN). The illustrative example - recognition and restoration of noisy voice pattern - was carried out using the TRHNN. The received TRHNN has not only high speed of signal processing, but also high speed of signal restoration. This example demonstrates that the proposed TRHNN is capable not only to identify the noisy input pattern, but also to restore the noise-free input pattern. The simulation results with use the software package Simulink show the computing procedure and applicability of TRHNNs for fast-acting signal processing and analysis in real-time conditions.
On-line references are not available - see the PDF file if available.
Faculty of Electrical Engineering and Computer Science
Stefan cel Mare University of Suceava, Romania
All rights reserved: Advances in Electrical and Computer Engineering is a registered trademark of the Stefan cel Mare University of Suceava. No part of this publication may be reproduced, stored in a retrieval system, photocopied, recorded or archived, without the written permission from the Editor. When authors submit their papers for publication, they agree that the copyright for their article be transferred to the Faculty of Electrical Engineering and Computer Science, Stefan cel Mare University of Suceava, Romania, if and only if the articles are accepted for publication. The copyright covers the exclusive rights to reproduce and distribute the article, including reprints and translations.
Permission for other use: The copyright owner's consent does not extend to copying for general distribution, for promotion, for creating new works, or for resale. Specific written permission must be obtained from the Editor for such copying. Direct linking to files hosted on this website is strictly prohibited.
Disclaimer: Whilst every effort is made by the publishers and editorial board to see that no inaccurate or misleading data, opinions or statements appear in this journal, they wish to make it clear that all information and opinions formulated in the articles, as well as linguistic accuracy, are the sole responsibility of the author.