Image Denoising via Clustering-Based Sparse Representation over Wiener and Gaussian Filters

In this paper, we present an adaptive denoising algorithm based on filter and sparse representation. It employs Wiener filter and Gaussian filter to extract high-frequency components on the noisy image, and simultaneously reduce the influence of noise for clustering. Image is denoised by solving a double-headed ℓ₁-optimization problem with the regularization involving dictionary learning and structural clustering. By conducting denoising on several commonly used images, the new algorithm performs equivalent and sometimes surpassing recently published leading algorithms in terms of both PSNR and visual quality.