Kidney Disease Classification from Ultrasound Images
A deep learning classifier for distinguishing normal, cyst, stone, and tumor conditions from kidney ultrasound images.
Problem Statement
Ultrasound-based kidney disease screening is cost-effective but subjective. Automated classification can assist radiologists, especially in resource-limited settings.
Method & Approach
Transfer learning with MobileNetV3 and EfficientNet-B0. Applied class-weighted loss to handle dataset imbalance. Evaluated with k-fold cross-validation.
Dataset / Data Source
Publicly available kidney ultrasound dataset (Kaggle: CT Kidney Dataset — adapted for ultrasound testing).
Results
Achieved over 93% accuracy on 4-class classification with consistent performance across cross-validation folds. Grad-CAM overlays highlight relevant kidney regions.
My Contribution
Preprocessing pipeline, model selection, training, evaluation, and visualization.