Scholarlink

COVID-19 pandemic has caused widespread fatalities and strained healthcare systems globally, prompting the need for efficient diagnostic tools. Scientists and healthcare professionals have sought solutions for predicting, analyzing, and detecting the virus. Advances in Machine Learning (ML) and Artificial Intelligence (AI) have enabled the development of systems to classify and predict COVID-19 cases using imaging data. This study uses deep learning, specifically Convolutional Neural Networks (CNNs), to distinguish CT scans of COVID-19 patients from non-COVID cases. Models like VGG19, DenseNet201, Xception, InceptionV3, ResNet101 and VGG16 were evaluated for performance. After 50 epochs, VGG16 and DenseNet201 achieved the highest accuracy of 98%, followed by VGG19, Xception, InceptionV3, and ResNet101 with accuracies of 97%, 96%, and 84%, respectively. The dataset included 400 normal and 399 COVID-19 CT scans. Image preprocessing techniques, such as contrast enhancement and noise reduction, improved image quality, optimizing model training and accuracy. Extending the training using 25 and 50 epochs further refined the model’s ability to learn complex patterns associated with COVID-19. In parallel, the development of a COVID-19 Prediction Web Application (CPWA) based on a fine-tuned, pre-trained VGG16 model underscores the real-world applicability of this approach, empowering healthcare professionals with an AI-driven tool for rapid, reliable screening. In summary, this study advances AI-driven medical imaging for COVID-19 diagnostics, offering a scalable solution to support healthcare professionals in combating the pandemic.

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

COVID-19 Prediction of Chest Computerized Tomography Images Using Convolutional Neural Networks

  • 会议论文

  • 发表日期：

    2025年

查看全文
  • Khokhoni Innocentia Mpho Ramaphosa,
  • Tranos Zuva,
  • Temidayo Otunniyi

摘要

COVID-19 pandemic has caused widespread fatalities and strained healthcare systems globally, prompting the need for efficient diagnostic tools. Scientists and healthcare professionals have sought solutions for predicting, analyzing, and detecting the virus. Advances in Machine Learning (ML) and Artificial Intelligence (AI) have enabled the development of systems to classify and predict COVID-19 cases using imaging data. This study uses deep learning, specifically Convolutional Neural Networks (CNNs), to distinguish CT scans of COVID-19 patients from non-COVID cases. Models like VGG19, DenseNet201, Xception, InceptionV3, ResNet101 and VGG16 were evaluated for performance. After 50 epochs, VGG16 and DenseNet201 achieved the highest accuracy of 98%, followed by VGG19, Xception, InceptionV3, and ResNet101 with accuracies of 97%, 96%, and 84%, respectively. The dataset included 400 normal and 399 COVID-19 CT scans. Image preprocessing techniques, such as contrast enhancement and noise reduction, improved image quality, optimizing model training and accuracy. Extending the training using 25 and 50 epochs further refined the model’s ability to learn complex patterns associated with COVID-19. In parallel, the development of a COVID-19 Prediction Web Application (CPWA) based on a fine-tuned, pre-trained VGG16 model underscores the real-world applicability of this approach, empowering healthcare professionals with an AI-driven tool for rapid, reliable screening. In summary, this study advances AI-driven medical imaging for COVID-19 diagnostics, offering a scalable solution to support healthcare professionals in combating the pandemic.