Patch-Based Image Denoising in the Wavelet Domain.
Bekır Dızdaroğlu · IPCV · 2010
Patch-based methods used in digital image processing fields are generally able to produce effective results. Although these approaches use easier structures to achieve better visual quality in digital image restoration compared with other methods, research is still going on in the field. In this study, a better noise reduction approach is presented using a patch-based algorithm in the wavelet domain. We implemented an isotropic patch-based smoothing method similar to the non-local means algorithm in the wavelet domain. Results of the method show that additive Gaussian noise is efficiently removed from grayscale images while image edges and textures remain unblurred.