Image denoising based on wavelet thresholding and Wiener filtering in the wavelet domain

Abstract

Wavelet transform has become a very important tool in the field of image denoising. A wavelet transform is a localised analysis of time (space) frequency. It uses a telescopic translation operation to gradually multi-scale refine the signal function and finally adapt to time frequency. One popular approach involves thresholding the wavelet coefficients by using the soft or hard threshold. Another method of image denoising is the Wiener filtering in the wavelet domain. In this study, Gaussian white noise has been added to two grey scale images and the two different denoising methods have been used. By comparing the performance of the two methods, it can be found that the Wiener filtering in the wavelet domain is more prowerful.

Keywords

• image denoising
• wavelet transforms
• Wiener filters
• white noise
• image denoising
• wavelet thresholding
• Wiener filtering
• wavelet domain
• wavelet transform
• time frequency
• telescopic translation operation
• multiscale refine
• wavelet coefficients
• soft threshold
• hard threshold
• grey scale images
• different denoising methods