Noise reduction with the use of fuzzy based weighted average filter and improved fractional calculus: A hybrid technique

Abstract

Medical images are having an important role for diagnosis of many diseases. Most of the medical images are inevitably affected by noises when they are being acquired, stored and transmitted. These medical images need to be free from noise for better diagnosis. Thus, despeckling methods plays a vital role in medical image analysis and diagnosis. In this paper, a hybrid noise reduction method using fuzzy weighted mean filter and improved fractional calculus is proposed and analyzed. Fuzzy based weighted mean filter with 3X3 kernel is applied on every pixel of an image and weights of pixel value is assigned by fuzzy logic, changing the central pixel value of the kernel with weighted mean value of all surrounding pixels in the mask. This process reduces the noise while preserving edges. Improved Gr¨unwald-Letnikov (G-L) fractional order differentiation filter is applied on the resultant image to get denoised image. To assess the quality of despeckled image, peak signal-to-noise ratio (PSNR), speckle suppression index, correlation coefficient and edge preservation index are considered. Simulation results demonstrate that the proposed hybrid despeckling method depicts better result than other methods considered for comparison. It maintains edges and other important details of an image while denoising.