Real Time Kidney Stone Detection System Using Yolov8 and Yolov11 On CT Images

Abstract

Kidney stones are a common medical problem that can cause severe pain and complications if not detected early [1],[2]. CT scan imaging is widely used for kidney stone detection because it provides clear and detailed images [3],[5]. However, manual analysis of CT images is time-consuming and depends on the experience of radiologists, which may lead to errors [1],[5].This project proposes a real-time kidney stone detection system using deep learning techniques. The system uses YOLOv8 and YOLOv11 models to automatically detect kidney stones from CT scan images [3],[4]. CT images are first preprocessed to improve quality and reduce noise, and then given as input to the models. The detected stones are highlighted using bounding boxes [3],[5]. Since YOLO models perform detection in a single stage, they are suitable for real-time applications. YOLOv8 provides fast detection with good accuracy, while YOLOv11 offers improved performance, especially for small or less visible stones[4],[6]. The models are evaluated using accuracy, precision, recall, F1-score, and mAP[3],[8]. The results show that both models can effectively detect kidney stones, with YOLOv11 giving slightly better accuracy and YOLOv8 performing faster. The proposed system reduces diagnosis time and supports doctors in making quick and accurate decisions [5].