Efficient Kidney Stone Classification Using a Spatial Attention Convolutional Neural Network

It is important to put kidney stones in the right group -whether it is a tumour, stone, cyst or normal so that they can be found and dealt with correctly. By looking at important parts of medical pictures, our model sorts kidney stones into different groups. We call this a new SA-CNN (Spatial Attention Convolutional Neural Network). Adding a way for spatial attention helps the model focus on areas that are important for diagnosis. This helps it group things better. An Fl-score of 98. 4% means it is very good, with an accuracy of 98. 7%, a precision of98.5%, a recall of98.3%, and a group that was chosen with care. It shows that our method can handle all kinds of kidney stones because it always works better than regular CNNs. If doctors use both deep learning and attention processes together, they might be able to make better guesses, look at data faster, and come up with more personalized treatment plans.