AI-Powered COVID-19 Detection: Researchers Test Modified Deep Learning Models

Strategic Insights -

Researchers at the University of Oklahoma have investigated the effectiveness of AI-powered models in detecting COVID-19 from chest X-ray images. The study, titled "Detecting Sars-cov-2 From Chest X-ray Using Artificial Intelligence," proposes and tests six modified deep learning models to identify patients with COVID-19 symptoms. The results, which are evaluated using a small and balanced dataset and a larger and imbalanced dataset, show promising accuracy rates, with some models achieving an accuracy of up to 100% in identifying COVID-19 patients. The researchers also present a pilot test of VGG16 models on a multi-class dataset, achieving 91% accuracy in detecting COVID-19, normal, and Pneumonia patients.

Key Takeaways:

  • The study proposes and tests six modified deep learning models, including VGG16, InceptionResNetV2, ResNet50, MobileNetV2, ResNet101, and VGG19, to detect SARS-CoV-2 infection from chest X-ray images.
  • Results show that VGG16 and MobileNetV2 models achieved an accuracy of up to 100% in identifying COVID-19 patients on both datasets, with a 95% confidence interval.
  • The study presents a pilot test of VGG16 models on a multi-class dataset, achieving 91% accuracy in detecting COVID-19, normal, and Pneumonia patients.
  • The researchers demonstrate that poorly performing models in Study One (ResNet50 and ResNet101) had their accuracy rise from 70% to 93% once trained with the larger dataset of Study Two.
  • The study shows that models like InceptionResNetV2 and VGG19's demonstrated an accuracy of 97% on both datasets, positing the effectiveness of the proposed methods.

Statistics:

  • The study achieved an accuracy rate of up to 100% in identifying COVID-19 patients using the VGG16 and MobileNetV2 models.
  • The study achieved 91% accuracy in detecting COVID-19, normal, and Pneumonia patients using the VGG16 model on a multi-class dataset.
  • The study demonstrated that poorly performing models in Study One (ResNet50 and ResNet101) had their accuracy rise from 70% to 93% once trained with the larger dataset of Study Two.
  • The study found that models like InceptionResNetV2 and VGG19's demonstrated an accuracy of 97% on both datasets.

Sources:

  • Detecting Sars-cov-2 From Chest X-ray Using Artificial Intelligence. IEEE Access, 2021;9:35501-35513. IEEE Access can be contacted at: Ieee-inst Electrical Electronics Engineers Inc, 445 Hoes Lane, Piscataway, NJ 08855-4141, USA.
  • NewsRx. Researchers from University of Oklahoma Provide Details of New Studies and Findings in the Area of Artificial Intelligence (Detecting Sars-cov-2 From Chest X-ray Using Artificial Intelligence). Robotics & Machine Learning. April 5, 2021; p 974.