A joint filtering based SVD technique for image de-noising

Image de-noising is an important part of image enhancement. This paper proposes an image de-noising technique on the singular value decomposition (SVD) using a combination of the signal-to-noise ratio (SNR) and median of noised image as a filtering function for singular values replacement. The signal noise is assumed to be an additive white Gaussian noise. Experimental results show that the proposed method is able to successfully enhance the visual quality of noisy images through this low complexity de-noising process.