Denoising CT Images using Median based Filters: a Review

Abstract

medical imaging is one of the essential tools for evidence-based medical diagnosis. However, salt and pepper noise could corrupt the original image, reducing the overall image quality. Computed tomography (CT) images database were used. The filter execution and evaluation algorithm were implemented using MATLAB environment. This article was conducted to study the performance of four different median based filters standard median filter (SMF), adaptive median filter AMF, center weight median filter (CWMF), and progressive switching median filter (PSMF), when applied to medical images. Noise immunity and edge-preserving were evaluated to characterizing the filtrations processes, by means of statistical (texture) and mathematical measures Peak Signal to Noise Ratio (PSNR), Mean Square Error (MSE), Correlation Ratio (CORR), and Image Enhancement Factor (IEF) for noise reduction, and automatic edge detection as visual evaluation for edges. The results shown that the Adaptive Median Filter(AMF) can remove the salt and pepper noise from CT image, the AMF algorithm maintain the edge of the image and detail information of the objects, And the overall filters comparison indicates a quite effective noise removal and satisfactory performance of AMF among others.