Image denoising using multiresolution principal component analysis

Abstract

Using Principal Component Analysis, noisy image is decorrelated so as to get distinction between signal and noise. Including the local characteristics of the image in the analysis procedure, retention of the important high frequency characteristics such as edges, curves, etc has become possible. For this purpose, a multiresolution environment is incorporated in the denoising process. By using the multiresolution image analysis, the human visual characteristics are maintained in the denoised image. Hence the image is more pleasing and informative.