Medical image enhancement using fractional derivatives

The present paper proposes a medical image enhancement scheme using a fractional differential function that selects the order of fractional derivative dynamically corresponding to each pixel of the image. The order depends upon the type of region to which the pixel belongs. The scheme incorporates the correlation between the neighboring pixels in cementing the mask to be applied to the image. The scheme focuses on enhancing the edges and texture while keeping the smooth regions of an image intact. Experimental results are obtained over a dataset of MRI images acquired from a diagnostic center. The comparison in terms of entropy, AMBE and PSNR clearly shows that the proposed method performs better than the other method.