Wavelet denoising in non gaussian noise using MDL principle

Wavelet methods have succeeded in image denoising in gaussian noise. In non-gaussian noise, however, these methods will degrade drastically. By employing the MDL principle and a wavelet coefficient model, this paper discusses a study on image denoising in gaussian mixture noise. A new denoising sche...

Full description

Saved in:
Bibliographic Details
Main Authors: Jiecheng Xie, Dali Zhang, Wenli Xu
Format: Conference Proceeding
Language:English
Subjects:
Active noise reduction
Array signal processing
Automation
Degradation
Gaussian noise
Image coding
Image denoising
Noise reduction
Wavelet coefficients
Wavelet domain
Online Access:Request full text
Tags: Add Tag
No Tags, Be the first to tag this record!
Description
Summary:Wavelet methods have succeeded in image denoising in gaussian noise. In non-gaussian noise, however, these methods will degrade drastically. By employing the MDL principle and a wavelet coefficient model, this paper discusses a study on image denoising in gaussian mixture noise. A new denoising scheme is derived and is based on per pixel detection. Experiment results show that the new scheme can not only denoise the image with small square error, but also provide a facility for the further compression.
DOI:10.1109/WCICA.2002.1021450