A Deep Learning Framework for Kidney Stone Prediction

Kidney stones are the most common chronic disease nowadays. Solid minerals that accumulate can lead to various sizes and shapes of stones inside the kidney. It requires innovative models to improve patient diagnosis with high precision, minimizing radiologist burden and giving them a tool that can automatically analyze kidney health, lowering the chance of misdiagnosis to obtain treatment and maintain a healthy lifestyle. The main aim of the research is to accurately identify kidney stones from computed tomography (CT) using Deep learning models. We presented effective and unique pre-trained deep learning models VGG16, ResNet50, and CNN for the identification of kidney stones accurately. The eminence of the adopted models has been validated in terms of positive, negative, and neutral measures. Moreover, the accuracy of the ResNet50 model is 7.7 and 4.5% superior to the VGG16 and CNN models, respectively.