Scholarlink

<p>The COVID-19 pandemic emerged as a severe and potentially fatal disease primarily affecting the lungs and posing a significant threat to human health. Chest X-ray and computed tomography (CT) imaging are commonly used for rapid and reliable medical diagnosis of COVID-19. The interpreting of these medical images is challenging as it requires constant attention and is prone to human error. In this paper, Detection and Predictive Analysis of COVID-19 based on Optimized Situational Aware Multi Graph Convolutional Recurrent Network (DPA-CD-SAMCRN) is proposed. At first, the input Images are gathered from CT Image dataset, Thermal Camera Image dataset and Audio dataset. Then, the collected images and signals are fed to preprocessing using Cross-Grained Neural Collaborative Filtering to resize the input image and remove unwanted noises from the signal data. Then, the preprocessed data is given into feature extraction stage with the help of Holistic Dynamic Frequency Transformer to extract gray scale statistical features, such as smoothness, energy, entropy, homogeneity. The extracted features are classified using Situational-Aware Multi-Graph Convolutional Recurrent Network (SAMCRN) for detecting COVID-19 infection as Positive or Negative, fever as Healthy or Unhealthy and Difficulty in breathing as Healthy or Unhealthy. To enhance accuracy, the Elite opposite Sparrow Search Optimization Algorithm is utilized to optimize SAMCRN parameters, ensuring precise detection. The performance measures like accuracy, precision, recall, f1-score, computational time and ROC are evaluated. The efficiency of the DPA-CD-SAMCRN achieves 16.84%, 18.65% and 12.17% higher accuracy; 10.24%, 13.89% and 23.16% higher precision 18.26%, 11.64% and 21.89% lower computational time compared with existing SCD-RNN-FEUL, SCD-RCNN-FKMC and SCC-DI-DCNN models respectively.</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

Detection and Predictive Analysis of COVID-19 Based on Optimized Situational Aware Multi Graph Convolutional Recurrent Network

  • 期刊论文

  • 发表日期：

    2026年1月29日

查看全文
  • V. Muthu,
  • S. Kavitha

摘要

The COVID-19 pandemic emerged as a severe and potentially fatal disease primarily affecting the lungs and posing a significant threat to human health. Chest X-ray and computed tomography (CT) imaging are commonly used for rapid and reliable medical diagnosis of COVID-19. The interpreting of these medical images is challenging as it requires constant attention and is prone to human error. In this paper, Detection and Predictive Analysis of COVID-19 based on Optimized Situational Aware Multi Graph Convolutional Recurrent Network (DPA-CD-SAMCRN) is proposed. At first, the input Images are gathered from CT Image dataset, Thermal Camera Image dataset and Audio dataset. Then, the collected images and signals are fed to preprocessing using Cross-Grained Neural Collaborative Filtering to resize the input image and remove unwanted noises from the signal data. Then, the preprocessed data is given into feature extraction stage with the help of Holistic Dynamic Frequency Transformer to extract gray scale statistical features, such as smoothness, energy, entropy, homogeneity. The extracted features are classified using Situational-Aware Multi-Graph Convolutional Recurrent Network (SAMCRN) for detecting COVID-19 infection as Positive or Negative, fever as Healthy or Unhealthy and Difficulty in breathing as Healthy or Unhealthy. To enhance accuracy, the Elite opposite Sparrow Search Optimization Algorithm is utilized to optimize SAMCRN parameters, ensuring precise detection. The performance measures like accuracy, precision, recall, f1-score, computational time and ROC are evaluated. The efficiency of the DPA-CD-SAMCRN achieves 16.84%, 18.65% and 12.17% higher accuracy; 10.24%, 13.89% and 23.16% higher precision 18.26%, 11.64% and 21.89% lower computational time compared with existing SCD-RNN-FEUL, SCD-RCNN-FKMC and SCC-DI-DCNN models respectively.