Image denoising using non-negative sparse coding shrinkage algorithm

This paper proposes a new method for denoising natural images using our extended non-negative sparse coding (NNSC) neural network shrinkage algorithm, which is self-adaptive to the statistic property of natural images. The basic principle of denoising using NNSC shrinkage is similar to that using st...

Full description

Saved in:
Bibliographic Details
Published in2005 IEEE Computer Society Conference on Computer Vision and Pattern Recognition (CVPR'05) Vol. 1; pp. 1017 - 1022 vol. 1
Main Authors Li Shang, Deshuang Huang
Format Conference Proceeding
LanguageEnglish
Published IEEE 2005
Subjects
Additive noise
Gaussian noise
Image coding
Image denoising
Neural networks
Noise level
Noise measurement
Noise reduction
Statistics
Testing
Online AccessGet full text

Cover

More Information
Summary:This paper proposes a new method for denoising natural images using our extended non-negative sparse coding (NNSC) neural network shrinkage algorithm, which is self-adaptive to the statistic property of natural images. The basic principle of denoising using NNSC shrinkage is similar to that using standard sparse shrinkage and wavelet soft threshold. Using test images corrupted by additive Gaussian noise, we evaluated the method across a range of noise levels. We utilized the normalized mean squared error as a measure of the quality of denoising images and the signal to noise rate (SNR) value as an evaluative feature of different denoising approaches. The experimental results prove that the NNSC shrinkage certainly is effective in image denoising. Otherwise, we also compare the effectiveness of the NNSC shrinkage with sparse coding shrinkage and wavelet soft threshold method. The simulative tests show that our denoising method outperforms any other of the two kinds of denoising approaches.
ISBN:0769523722
9780769523729
ISSN:1063-6919
1063-6919
DOI:10.1109/CVPR.2005.183