An Automated System of Identifying Kidney Stones using Darknet-19, a Deep Learning Model in MATLAB.

Abstract

The excretory system depends on the kidneys, which are essential organs. The kidney is in charge of eliminating waste products from the body through urination. Kidney stones are currently the most frequent issue a person has with their kidneys. One in ten persons will receive a kidney stone diagnosis at some point in their lives. Hard deposits of calcium, salt, or other minerals that are not effectively eliminated through urine are known as kidney stones. Most kidney stones form in the region of the urinary tract. Kidney stones can be found using a variety of diagnostic techniques. Despite the fact that doctors have trouble detecting tiny stones, their position and size. In most of the hospitals, the size of kidney stones is detected manually. In order to reduce false negative results, improving diagnosis and to help the radiologists in identifying the accurate problem, we proposed an automated method to detect kidney stones using a deep learning model. In our project, we use darknet 19 (a deep learning model) for training the datasets and for feature extraction purpose. Using the extracted features, we can classify the images for predicting the accurate result. We can predict whether the kidney image is normal or abnormal, whether there is presence of kidney stones, kidney stone size and their location are found.