Enhancing Kidney Stone Diagnosis: A Fusion Approach of FCM and CNN for Precise Detection

Kidney stones, characterized by the accumulation of mineral and organic substances within the renal system, pose a significant health concern globally. Timely and accurate diagnosis is paramount for effective treatment and prevention of complications. Medical imaging, particularly computed tomography (CT) scans, plays a crucial role in identifying and characterizing kidney stones. However, the inherent variability in stone appearance, shape, and size, coupled with the potential presence of artifacts in imaging data, poses challenges to traditional diagnostic methods. Hence, this work focuses on addressing these challenges and dedicated to optimizing the detection of kidney stones by employing a synergistic approach that integrates image processing and deep learning techniques. The proposed work involves initial pre-processing steps, utilizing median filtering to eliminate noise and adaptive histogram equalization (AHE) for enhancing the quality of CT images of kidneys. Subsequently, convolutional neural networks (CNNs) are employed as a deep learning algorithm for classification, with a primary focus on accurately distinguishing between normal and abnormal images. In the event of an abnormal classification indicating the presence of a kidney stone, a secondary step is implemented. This step involves the application of fuzzy C-means (FCM) and level set segmentation techniques to achieve precise detection and localization of the kidney stone within the image. The proposed approach achieves an accuracy of 99.45%, which is 13% improvised when compared to the existing approaches.