Chronic kidney disease (CKD) is a widespread medical issue that results from diminished kidney function and, in severe cases, leads to kidney failure. One of the contributing factors to impaired kidney performance is the development of kidney stones. Since this condition often shows no noticeable symptoms during its early stages, timely and accurate diagnosis is crucial to preventing serious complications. In this research, we propose a highly effective and reliable method for detecting kidney stones by leveraging ensemble deep learning models enhanced through inductive transfer learning. The methodology incorporates data from two main sources: the Kidney Data and the CT Kidney Stone datasets. For classification purposes, a variety of deep learning architectures were utilized, including DarkNet19, InceptionV3, ResNet101, DenseNet169, MobileNetV2, VGG16, GoogleNet, AlexNet, ShuffleNet, SqueezeNet, and a custom-designed DNN model (FindWell). These models were further supported by feature extraction and selection processes using the ReliefF algorithm. Classification accuracy was validated through K-Nearest Neighbors (KNN) and K-Fold cross-validation techniques. For the detection task, different models from the YOLO (You Only Look Once) family—namely YOLO v5x6, v5s6, v8n, and v9n—were deployed to identify kidney stones in imaging data. In addition, the Xception architecture was applied for a comprehensive analysis of the dataset. This integrated approach of combining multiple cutting-edge algorithms enhances both the precision and speed of kidney stone identification, which can significantly aid in the early diagnosis and treatment of patients suffering from chronic kidney-related disorders.
Introduction
Kidney diseases affect people of all ages and early detection is crucial to prevent serious complications such as chronic kidney disease and worsening conditions from kidney stones. The global rise in kidney disease cases, especially in developing countries with few nephrologists, creates a need for efficient diagnostic tools. Traditional screening and imaging methods are often time-consuming and prone to human error.
Automated systems using machine learning (ML) and deep learning (DL) have been developed to assist in early and accurate identification of kidney disorders. These tools reduce workload on healthcare providers, improve diagnostic accuracy, and enable faster, more objective medical decisions. Various studies have demonstrated the effectiveness of ML models—such as ensemble learning, radiomics, and traditional classifiers like KNN and SVM—in classifying kidney stones and differentiating related conditions.
The proposed system uses an ensemble of deep neural networks combined with advanced feature selection techniques for both classification and detection of kidney stones from medical images. It leverages multiple state-of-the-art CNN architectures (e.g., DarkNet19, ResNet101, DenseNet169, MobileNetV2) alongside YOLO models (YOLOv5 to YOLOv9) for precise detection. The system uses two main datasets—Kidney Data and CT Kidney Stone Data—and applies extensive preprocessing, including image augmentation, to improve training.
The approach integrates feature extraction (using CNN and HOG methods), optimized feature selection (ReliefF), and K-Nearest Neighbors with K-Fold validation to boost accuracy. The detection pipeline converts images into annotated blob objects and applies bounding box techniques. This comprehensive framework aims to offer a scalable, reliable solution for early kidney stone detection, enhancing diagnostic performance and potentially reducing kidney disease progression risks.
Conclusion
This project successfully demonstrates the effectiveness of advanced deep learning methods for detecting and classifying kidney stones. By applying models such as DarkNet19, ResNet101, DenseNet169, and Xception, the system was able to achieve strong performance in analyzing CT scan images. These deep learning architectures are capable of extracting significant features from medical images, which allows accurate identification of different types and sizes of kidney stones. The integration of YOLOv5 and YOLOv8 further strengthened the system by enabling real-time detection and precise localization of abnormalities. Techniques such as data augmentation, feature extraction, and ensemble learning played a key role in improving both accuracy and reliability. Altogether, the proposed system provides a holistic approach to tackling the challenge of kidney stone diagnosis. This project emphasizes the role of advanced technologies in supporting healthcare professionals and improving patient care.
References
[1] K. K. Patro, J. P. Allam, B. C. Neelapu, R. Tadeusiewicz, U. R. Acharya, M. Hammad, O. Yildirim, and P. Plawiak, ‘‘Application of Kronecker convolutions in deep learning technique for automated detection of kidney stones with coronal CT images,’’ Inf. Sci., vol. 640, Sep. 2023, Art. no. 119005, doi: 10.1016/j.ins.2023.119005.
[2] O. Sabuncu, B. Bilgehan, E. Kneebone, and O. Mirzaei, ‘‘Effective deep learning classification for kidney stone using axial computed tomography (CT) images,’’ Biomed. Eng./Biomedizinische Technik, vol. 68, no. 5, pp. 481–491, 2023, doi: 10.1515/bmt-2022-0142.
[3] C. Yan and N. Razmjooy, ‘‘Kidney stone detection using an optimized deep believe network by fractional coronavirus herd immunity optimizer,’’ Biomed. Signal Process. Control, vol. 86, Sep. 2023, Art. no. 104951, doi: 10.1016/j.bspc.2023.104951.
[4] S. S. Vidhya, D. Vishmitha, K. Yoshika, P. Sivalakshmi, V. Chowdary, K. G. Shanthi, and M. Yamini, ‘‘Kidney stone detection using deep learning and transfer learning,’’ in Proc. 4th Int. Conf. Inventive Res. Comput. Appl. (ICIRCA), Sep. 2022, pp. 987–992, doi: 10.1109/ICIRCA54612.2022.9985723.
[5] S. Bhardwaj and H. S. Shreenidhi, ‘‘Modeling of an CNN architecture for kidney stone detection using image processing,’’ in Proc. 4th Int. Conf. Emerg. Res. Electron., Comput. Sci. Technol. (ICERECT), Dec. 2022, pp. 1–5, doi: 10.1109/ICERECT56837.2022.10059972.
[6] National Kidney Foundation. Accessed: Jan. 2, 2023. [Online]. Available: https://www.kidney.org/kidneydisease/global-facts-about-kidney-disease
[7] E. Jussupow, K. Spohrer, A. Heinzl, and J. Gawlitza, ‘‘Augmenting medical diagnosis decisions? An investigation into physicians’ decisionmaking process with artificial intelligence,’’ Inf. Syst. Res., vol. 32, no. 3, pp. 713–735, Sep. 2021, doi: 10.1287/isre.2020.0980.
[8] V. Singh, V. K. Asari, and R. Rajasekaran, ‘‘A deep neural network for early detection and prediction of chronic kidney disease,’’ Diagnostics, vol. 12, no. 1, p. 116, Jan. 2022, doi: 10.3390/diagnostics12010116.
[9] I. Castiglioni, L. Rundo, M. Codari, G. Di Leo, C. Salvatore, M. Interlenghi, F. Gallivanone, A. Cozzi, N. C. D’Amico, and F. Sardanelli, ‘‘AI applications to medical images: From machine learning to deep learning,’’ Phys. Medica, vol. 83, pp. 9–24, Mar. 2021, doi: 10.1016/j.ejmp.2021.02.006.
[10] M. Biswas, V. Kuppili, L. Saba, D. R. Edla, H. S. Suri, E. Cuadrado-Godia, J. R. Laird, R. T. Marinhoe, J. M. Sanches, A. Nicolaides, and J. S. Suri, ‘‘State-of-the-art review on deep learning in medical imaging,’’ Frontiers Biosci., vol. 24, no. 3, pp. 392–426, 2019, doi: 10.2741/4725.
[11] H. E. Kim, A. Cosa-Linan, N. Santhanam, M. Jannesari, M. E. Maros, and T. Ganslandt, ‘‘Transfer learning for medical image classification: A literature review,’’ BMC Med. Imag., vol. 22, no. 1, p. 69, Dec. 2022, doi: 10.1186/s12880-022-00793-7.
[12] S. T. Krishna and H. K. Kalluri, ‘‘Deep learning and transfer learning approaches for image classification,’’ Int. J. Recent Technol. Eng. (IJRTE), vol. 7, pp. 427–432, Feb. 2019.
[13] N. Aslan, G. Ozmen Koca, M. A. Kobat, and S. Dogan, ‘‘Multiclassification deep CNN model for diagnosing COVID-19 using iterative neighborhood component analysis and iterative ReliefF feature selection techniques with X-ray images,’’ Chemometric Intell. Lab. Syst., vol. 224, May 2022, Art. no. 104539, doi: 10.1016/j.chemolab.2022.104539.
[14] L. D. Sharma, H. Chhabra, U. Chauhan, R. K. Saraswat, and R. K. Sunkaria, ‘‘Mental arithmetic task load recognition using EEG signal and Bayesian optimized K-nearest neighbor,’’ Int. J. Inf. Technol., vol. 13, no. 6, pp. 2363–2369, Dec. 2021, doi: 10.1007/s41870-021-00807-7.
[15] J. Serrat, F. Lumbreras, F. Blanco, M. Valiente, and M. López-Mesas, ‘‘MyStone: A system for automatic kidney stone classification,’’ Exp. Syst. Appl., vol. 89, pp. 41–51, Dec. 2017, doi: 10.1016/j.eswa.2017.07.024.
[16] A. Caglayan, M. O. Horsanali, K. Kocadurdu, E. Ismailoglu, and S. Guneyli, ‘‘Deep learning model-assisted detection of kidney stones on computed tomography,’’ Int. Braz. J., vol. 48, no. 5, pp. 830–839, Oct. 2022, doi: 10.1590/s1677-5538.ibju.2022.0132.
[17] J. Verma, M. Nath, P. Tripathi, and K. K. Saini, ‘‘Analysis and identification of kidney stone using kth nearest neighbour (KNN) and support vector machine (SVM) classification techniques,’’ Pattern Recognit. Image Anal., vol. 27, no. 3, pp. 574–580, Jul. 2017, doi: 10.1134/s1054661817030294.
[18] T. De Perrot, J. Hofmeister, S. Burgermeister, S. P. Martin, G. Feutry, J. Klein, and X. Montet, ‘‘Differentiating kidney stones from phleboliths in unenhanced low-dose computed tomography using radiomics and machine learning,’’ Eur. Radiol., vol. 29, no. 9, pp. 4776–4782, Sep. 2019, doi: 10.1007/s00330-019-6004-7.
[19] I. Aksakalli, S. Kaçdioglu, and Y. S. Hanay, ‘‘Kidney X-ray images classification using machine learning and deep learning methods,’’ Balkan J. Electr. Comput. Eng., vol. 9, no. 2, pp. 144–151, Apr. 2021, doi: 10.17694/bajece.878116.
[20] B. Manoj, N. Mohan, and S. Kumar, ‘‘Automated detection of kidney stone using deep learning models,’’ in Proc. 2nd Int. Conf. Intell. Technol. (CONIT), Jun. 2022, pp. 1–5, doi: 10.1109/CONIT55038.2022.9847894.