Scholarlink

Kidney stone disease is a one of the most common illnesses that affects kidney functionality and can cause to kidney failure. It is also recognized as global health problem affected by both men and women. Regular checkups like CT scans, X-Rays and Ultra Sound rays may help to identify the abnormality early and may reduce the mortality rates. This research presents a novel deep learning-based CNN model utilizing an automated detection tool on Computed Tomography Images of Kidney Stone. This proposed methodology follows various steps to detect and classify the kidney stone disease by following the pipeline as follows: Collecting CT-Scan Images, Image preprocessing and Data Augmentation, Model Training, Classification and Evaluation. The dataset is collected from Kaggle and undergoes data augmentation for improving the training process and preventing over-fitting. It further, applies various image preprocessing techniques for feeding the CNN model. Then the CNN model is trained by adapting the benefits of Adam optimizer for fine-tuning the hyper-parameters. Finally, the performance of the CNN model can be evaluated using the metrics of accuracy, precision, recall and F-score values. The proposed CNN model exhibits superior results with 0.9986 of accuracy when compared with other deep learning models like ResNet50 and EfficientNetB0.

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

CNN-Based Deep Learning Model for Kidney Stone Prediction Using CT Scan Images

  • 会议论文

  • 发表日期：

    2026年

查看全文
  • Ch V. S. Satyamurty,
  • Methari Sandhya Jayachandra

摘要

Kidney stone disease is a one of the most common illnesses that affects kidney functionality and can cause to kidney failure. It is also recognized as global health problem affected by both men and women. Regular checkups like CT scans, X-Rays and Ultra Sound rays may help to identify the abnormality early and may reduce the mortality rates. This research presents a novel deep learning-based CNN model utilizing an automated detection tool on Computed Tomography Images of Kidney Stone. This proposed methodology follows various steps to detect and classify the kidney stone disease by following the pipeline as follows: Collecting CT-Scan Images, Image preprocessing and Data Augmentation, Model Training, Classification and Evaluation. The dataset is collected from Kaggle and undergoes data augmentation for improving the training process and preventing over-fitting. It further, applies various image preprocessing techniques for feeding the CNN model. Then the CNN model is trained by adapting the benefits of Adam optimizer for fine-tuning the hyper-parameters. Finally, the performance of the CNN model can be evaluated using the metrics of accuracy, precision, recall and F-score values. The proposed CNN model exhibits superior results with 0.9986 of accuracy when compared with other deep learning models like ResNet50 and EfficientNetB0.