A single natural image denoising based on independent component analysis

Abstract: Denoising of natural image is challenging problem in image processing. Image denoising problem is very simple. Independent component analysis (ICA) is higher order statistical tool for analysis of multidimensional data. Image denoising using independent component analysis assume an Gaussian noise. The noise is a Gaussian and considered image data is non-Gaussian random variable. In practical situations type of noise, type of image, amount of noise are all variable parameters. Such a parameter cannot effectively remove the noise. In this paper we use a independent component analysis to remove a noise from a single image. Natural images provide the basic knowledge for understanding.

Keywords: Adaptive PCA, Denoise, ICA, PCA.