Medical image denoising using multistage directional median filter

In modern medical field visual information transmitted in form of digital images is becoming a inevitable method of communication. The choice of denoising algorithm is application dependant and depends on the noise added in image. In this paper multistage median filtering (MMF) is used which is an extension of median filter that uses four directional windows.

MMF provides spatial structural characteristics which can preserve edges effectively by differentiating between flat region and edge region using the maximum and minimum median value. Performance is evaluated for MMF using each of the windows and their combination. The indexes Peak Signal to Noise Ratio (PSNR) and Root Mean Square Error (RMSE) demonstrate marked improvement of MMF over wavelet method.