Scholarlink

<p>Kidney disease is a major health concern worldwide, requiring exact identification for giving proper treatment. Deep learning technologies are widely used as important tools in the biomedical industry for disease detection. The modern deep network techniques frequently experience overfitting and low accuracy issues, requiring additional changes for optimum performance. In this paper, an Automated Kidney Stone Detection and Classification using Dynamic Generative Residual Graph Convolutional Neural Networks for Computed Tomography-Based Medical Imaging (KSDC-DGRGCNN-CT-MI) is proposed. The input image is collected from CT Kidney dataset. Then, the collected images are pre-processing by using Regularized Bias aware Ensemble Kalman Filtering (RBEKF) for noise reduction, contrast enhancement, and resizing. The pre-processed images are given into the feature extraction using Multi-Hypothesis Fuzzy Matching Radon Transform (MHFMRT) to extract texture features, such as entropy, contrast, and correlation. The extracted features are fed into the Kidney Stone Detection using Dynamic Generative Residual Graph Convolutional Neural Networks (DGRGCNN) to detect and categorize the disease as Cyst, Normal, Stone, and Tumor. Finally, Superb Fairy-wren Optimization Algorithm (SFOA) is utilized for optimizing DGRGCNN, which accurately detect kidney stone. The proposed KSDC-DGRGCNN-CT-MI method is implemented and the effectiveness is assessed under some metrics, such as accuracy, precision, recall, and computational time. The proposed KSDC-DGRGCNN-CT-MI method attains 99.12% higher accuracy and 98.01% higher precision compared with existing methods.</p>

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

Automated Kidney Stone Detection and Classification utilizing Dynamic Generative Residual Graph Convolutional Neural Networks for Computed Tomography-Based Medical Imaging

  • 期刊论文

  • 发表日期：

    2025年12月29日

查看全文
  • Shivani Verma,
  • Pawan Kumar Singh

摘要

Kidney disease is a major health concern worldwide, requiring exact identification for giving proper treatment. Deep learning technologies are widely used as important tools in the biomedical industry for disease detection. The modern deep network techniques frequently experience overfitting and low accuracy issues, requiring additional changes for optimum performance. In this paper, an Automated Kidney Stone Detection and Classification using Dynamic Generative Residual Graph Convolutional Neural Networks for Computed Tomography-Based Medical Imaging (KSDC-DGRGCNN-CT-MI) is proposed. The input image is collected from CT Kidney dataset. Then, the collected images are pre-processing by using Regularized Bias aware Ensemble Kalman Filtering (RBEKF) for noise reduction, contrast enhancement, and resizing. The pre-processed images are given into the feature extraction using Multi-Hypothesis Fuzzy Matching Radon Transform (MHFMRT) to extract texture features, such as entropy, contrast, and correlation. The extracted features are fed into the Kidney Stone Detection using Dynamic Generative Residual Graph Convolutional Neural Networks (DGRGCNN) to detect and categorize the disease as Cyst, Normal, Stone, and Tumor. Finally, Superb Fairy-wren Optimization Algorithm (SFOA) is utilized for optimizing DGRGCNN, which accurately detect kidney stone. The proposed KSDC-DGRGCNN-CT-MI method is implemented and the effectiveness is assessed under some metrics, such as accuracy, precision, recall, and computational time. The proposed KSDC-DGRGCNN-CT-MI method attains 99.12% higher accuracy and 98.01% higher precision compared with existing methods.