A hybrid of deep and textural features to differentiate glomerulosclerosis and minimal change disease from glomerulus biopsy images

The minimal change disease (MCD) and glomerulosclerosis (GS) are two common kidney diseases. Unless adequately treated, these diseases leads to chronic Accurate differentiation of is paramount importance as their methods treatment prognoses different. Thus, this article propose a method capable differentiating MCD from GS in glomerulus biopsies images based on new hybrid deep texture feature space. We conducted an extensive study determine the best set features for image representation. Our extraction methodology, which includes Haraliks geostatistics descriptors pre-trained CNNs, resulted 13,476 characteristics. then used mutual information order elements by select using random forest classifier. proposed achieved accuracy 90.3% Kappa index 80.5%. Representation biopsy with textural facilitates accurate MCD.