Deep Learning-Based Automated Diagnosis of Kidney Stones

Kidney stone disease is a widespread medical condition that requires timely and accurate detection for effective treatment. This study proposes a deep learning-based approach for detecting kidney stones in CT scan images using convolutional neural networks (CNN). The dataset, consisting of labeled CT scans, is split into 70%-20%-10% for training, validation, and testing, respectively. Transfer learning is employed using the ResNet50 model, pre-trained on ImageNet, as a feature extractor. The classification head is modified to address the specific task of kidney stone detection. To enhance performance, fine-tuning techniques, adaptive learning rate scheduling, and data augmentation are implemented. The model’s accuracy and loss are evaluated on validation and test sets to ensure its robustness and generalizability. Additionally, a chatbot powered by a large language model (LLM) is integrated to assist patients by providing medical guidance and suggesting nearby hospitals for immediate consultation. The final trained model is stored in TensorFlow saved model for deployment. This study highlights the potential of AI-driven medical imaging solutions, offering a reliable and efficient tool to assist healthcare professionals in kidney stone diagnosis.