Deep Learning Based COVID-19 Detection via Hard Voting Ensemble Method

Area of Science:

• Medical Imaging
• Artificial Intelligence in Healthcare
• Infectious Disease Diagnostics

Background:

• Global healthcare systems face immense pressure due to the COVID-19 pandemic.
• Early and accurate diagnosis is crucial for controlling virus spread and patient treatment.
• Medical imaging, particularly X-rays, offers valuable insights into lung conditions.

Purpose of the Study:

• To develop and evaluate a novel ensemble approach for COVID-19 identification using X-ray images.
• To leverage deep learning and transfer learning for improved diagnostic performance.
• To enhance the efficiency of COVID-19 diagnosis in resource-constrained settings.

Main Methods:

• An ensemble method combining confidence scores from CNN, VGG16, and DenseNet models using hard voting.
• Application of transfer learning to optimize performance on limited medical image datasets.
• Utilizing X-ray-PIC (X-ray Pictures) for COVID-19 detection.

Main Results:

• The proposed ensemble approach achieved 97% accuracy, 96% precision, 100% recall, and 98% F1-score.
• Demonstrated superior performance compared to existing diagnostic techniques.
• Validated the effectiveness of transfer learning in enhancing model performance.

Conclusions:

• Ensemble methods combined with transfer learning show significant promise for COVID-19 diagnosis via X-ray.
• The X-ray-PIC approach can substantially aid in early disease detection.
• This methodology has the potential to alleviate the burden on global healthcare systems.