Enhancement of MRI images of brain tumor using Grunwald Letnikov fractional differential mask

The present paper focuses on the enhancement of magnetic resonance imaging (MRI) images of the brain tumor using the Grunwald Letnikov (G-L) fractional differential mask. The method aims to enhance the edges and texture while preserving the smooth regions of an image. This will help the doctors to take a right decision for treatment by correctly identifying the location of the tumor present in an image. The method uses the G-L definition of the fractional derivative to form masks of size 3 x 3 and 5 x 5 in which the correlation of the neighboring pixels is preserved. A gradient is used to find the threshold so that the input image can be partitioned into edge, texture and smooth region. The order of the fractional derivative is chosen individually for each pixel of these three regions and the framed mask is applied on the input image to get the enhanced image. To show the effectiveness of the proposed method, results are presented in terms of visual appearance, subjective assessment and quantitative metrics. PSNR, AMBE, entropy, and GLCM are used as evaluation parameters for quantitative analysis. The comparison with other existing methods such as fixed order fractional differential, adaptive fractional differential, and modified G-L differential operator shows the improvement in results obtained by the proposed method.