SEN-D

SEN-D detects kidney stones in CT scans using a stacked ensemble of three convolutional networks (InceptionV3, InceptionResNetV2 and Xception) combined by a meta-learner. It reaches 98.74% accuracy, and shows Grad-CAM heatmaps so you can see which part of the scan drove each decision. The model runs behind a FastAPI backend with a React frontend.