Noise Removal In MR Images Using Non-Linear Filters

Images are imitations of real world objects. Often an image is a two dimensional (2D) signal f(x,y) represent the amplitude or intensity of the image. In the Transmission of images, they are corrupted by salt and pepper noise, due to faulty communications. Salt and Pepper noise is also known as Impulse noise. The target of straining is to eliminate the compulsions so that the noise less image is fully improved with slightest signal alteration. The best-known and most commonly used nonlinear digital filters, based on order statistics are median filters, also known as Simple Median Filter (SMF). Median filters are acknowledged for their competency to eliminate impulse noise deprived of damaging the edges. Median filters are documented for their competency to confiscate impulse noise as well as preserve the boundaries. The effective confiscation of impulse often hints to images with distorted and inaccurate features. Ideally, the filtering should be applied only to corrupted pixels while leaving uncorrupted pixels intact. Applying median filter unconditionally across the entire image as practiced in the conventional schemes would inevitably alter the intensities and remove the signal details of uncorrupted pixels. Hence, a noise-detection process to categorize between virtuous pixels and the degraded pixels prior to smearing nonlinear filtering is highly desirable. The main aim of this work is to modify the existing median filters and implement the modified median filter for reduction of high density impulse noise (salt & pepper noise). Then evaluate the performance of the algorithm using MSE & PSNR parameters.