Prediction Of Chronic Kidney Disease On Ct Images Using Deep Learning: Resnet-34 And Vggnet-16

Kidney is considered as the most essential and important organ of our human body. Many factors such as diabetes, high blood pressure, inherited kidney disease, abnormal structure of kidney etc., affects the filtration rate of the kidney and in turn leads to chronic kidney disease. İf disease is identified and properly treated at early stages, human life can be saved. But in most cases, it cannot be done since there is no proper symptoms in the early stages. Huge of data is produced in the medical industry daily and it is left unattended. If this data is utilized properly, many useful information can be found. The use of Artificial Intelligence to assist medical professionals is increasing day by day. In this research work, we have used two famous deep learning architectures: RESNET and VGGNET to predict the existence of Chronic Kidney Disease. To improve the prediction accuracy, Artificial Bee Colony algorithm is used for segmentation. Instead of giving the original image as input, segmented image is given. The experiments shows that the VGGNET architecture produces higher accuracy of 98% than the RESNET architecture which gives an accuracy of 96%.