Denoising of MRI images

Magnetic Resonance (MR) images are corrupted with complex white Gaussian noise that is known as Rician noise. This noise is multiplicative in nature because it is signal-dependent. This noise significantly reduces the contrast quality of MR images. Therefore, a lot of difficulties have been faced by analysts to extract the useful information from images for proper diagnosis of disease. Various authors have proposed different spatial and transform based techniques for MR image denoising. Our main aim is to preserve useful details such as texture information, thin edges, and structural content of useful information when denoising criteria is applied to the image corrupted with Rician noise. In this work, we analyze the various spatialbased techniques, Spatial domain filtering techniques use local data redundancy to remove noise from image and they use edge detection techniques to recognize edges in the image. These techniques have ability not only to preserve the edges of fine structures but also to perform efficient denoising. All the experiments have been conducted on standard clinical data sets and efficiency of the various methods is measured by various qualitative and quantitative measures.