MRI denoising using non-local PCA with DWT

Abstract

Removing noise from original image is fundamental issue of image denoising. There are number of existing image denoising methods. Each method follows its own unique approach and each method has its own merits and limitations. In this paper, denoising of MR images (Magnetic Resonance Imaging) problem is solved by using a combined approach of non local Principal Component Analysis (PCA) and Discrete Wavelet Transform (DWT). Non local principal component analysis follows a patch based approach to remove the noise then discrete wavelet transform is applied to further improve. denoised image to get more denoised MR image. The results reveal that the proposed algorithm provides better peak signal, to noise ratio, structural similarity index and some other parameters along with more effectively denoising the MR images that also preserve the details of original image better. The results are calculated in terms of peak signal to noise ratio (PSNR), structural similarity index (SSIM), image enhancement, factor (IEF), correlation coefficient (CC), root mean square error (RMSE) and mean square error (MSE).