Document Details

Document Type : Article In Conference 
Document Title :
Improving the Recognition of Noisy Arabic Speech Via Wavelet Compression
تحسين عملية التعرف على الكلام العربي المشوش عن طريق ضغط المويجات
 
Document Language : English 
Abstract : This paper presents a method that uses the wavelet transform in compressing a noisy speech signal and then, testing its effect on the reconstruction of noisy Arabic speech utterances after decompression. The ability of wavelets in compacting signal energy is employed to separate the significant signal features from the noise contribution. The technique showed an improved SNR for the processed speech samples. In each experiment performed, the correlation coefficients are computed and compared to the correlation coefficient yielded by a similar compression scheme based on the Fourier transform. The use of the wavelet-based compression technique is superior in improving the recognition of speech signals especially at high input SNR values 
Conference Name : The 17th National Conference for Computers 
Duration : From : 15 صفر- 1425 AH - To : 18صفر- 1425 AH
From : 5 أبريل - 2004 AD - To : 8 أبريل - 2004 AD 
Publishing Year : 1425 AH
2004 AD 
Number Of Pages : 8 
Article Type : Article 
Conference Place : Al-Madinah Al-Munawwarah, Saudi Arabia 
Organizing Body : KING ABDULAZIZ UNIVERSITY 
Added Date : Saturday, June 23, 2012 

Researchers

Researcher Name (Arabic)Researcher Name (English)Researcher TypeDr GradeEmail
عبدالله احمد باسهيلBasuhail, Abdullah AhmadInvestigatorDoctorateabasuhail@yahoo.com

Files

File NameTypeDescription
 33730.pdf pdf 

Back To Researches Page