YOLOv8-Powered Multi-Class Kidney Abnormality Detection via CT Imaging

Kidney-related disorders affect individuals across all age groups and have become a significant public health concern due to the increasing prevalence of chronic kidney disease (CKD) and the global shortage of nephrologists. CKD is a progressive condition that deteriorates kidney function and often goes undetected in its early stages. Various factors, including tumors, cysts, and kidney stones, contribute to renal failure. Tumors can cause direct damage to surrounding organs, while kidney stones, formed by solid deposits, lead to painful blockages in the urinary system. Early detection and timely intervention are crucial in preventing severe complications. This study introduces an AI-powered diagnostic system for detecting kidney abnormalities in computed tomography (CT) scans. Utilizing the YOLOv8 deep learning model, the system was trained on a dataset of 12,446 CT images, classified into four categories: normal, cyst, tumor, and stone. The dataset was sourced from multiple hospitals in Dhaka, Bangladesh. The model's performance was assessed using key evaluation metrics, achieving an accuracy of 99.78%, precision of 85.76%, recall of 75.28%, F1-score of 99.78%, and specificity of 93.12%. These findings demonstrate the system’s potential to assist healthcare professionals by automating kidney abnormality detection, facilitating early diagnosis, and reducing the workload on medical practitioners.