A denoising approach via wavelet domain diffusion and image domain diffusion

This paper presents a new image denoising algorithm based on wavelet transform and nonlinear diffusion. Although the wavelet domain diffusion methods are very effective in image denoising, the salient artifacts are still produced. On the other hand, the image domain diffusion methods output the denoised image with fewer artifacts. So, unlike the previous denoising methods employing wavelet transform and diffusion scheme, the proposed method implements the diffusion not only in the wavelet domain but also in the image domain. The new method is called image denoising method combining the wavelet domain diffusion and the image domain diffusion (WDD-IDD). In the process of denoising, the initial denoised image is obtained by carrying out the wavelet domain isotropic diffusion. And then, the final denoised image is produced by applying the image domain anisotropic diffusion on the initial denoised image. It is noted that the image domain anisotropic diffusion scheme is constructed based on the feature of initial denoised image. In addition, to exemplify the power of the proposed method, the processing is restricted to the non-subsampled shearlet transform (NSST) domain which can better capture the geometry features of image than other wavelet transform. The tests show that the proposed WDD-IDD produces a better result in term of peak signal-to-noise ratio (PSNR), structural similarity (SSIM), and visual effect compared to the related methods.