Anisotropic diffusion-based denoising using residual image for preservation of image details

This paper presents an anisotropic diffusion-based approach to noise reduction, which utilizes the residual image in the wavelet domain to improve the quality of image detail preservation. In the experimental results, the proposed method improves the quality of denoised images by increasing peak signal-to-noise ratio up to 1.17 dB and structural similarity index up to 0.026.