Using Clinical Data and Deep Features in Renal Pathologies Classification

The identification and categorization of glomerular diseases are crucial in the diagnosis of numerous renal diseases. This research carried out a comprehensive investigation to establish the optimal group of characteristics for depicting glomerular pathological images. Our feature extraction methodology includes clinical data and deep features. In addition, we compared four classifiers to support the specialist defining a renal pathology diagnosis. The study found that by combining clinical information with deep features, a high level of accuracy (95.98%) and agreement (98.93%) was achieved using the Random Forest classifier. This suggests that the combination of these two types of data makes it easier to correctly classify different diseases.