COVID-19 disease severity assessment using CNN model.

Area of Science:

• Medical Imaging
• Artificial Intelligence
• Infectious Diseases

Background:

• The COVID-19 pandemic has overwhelmed healthcare systems, necessitating efficient patient stratification.
• Limited clinical resources require accurate assessment of disease severity for effective patient management.
• Chest X-rays are crucial for diagnosing COVID-19 but manual interpretation can be time-consuming.

Purpose of the Study:

• To develop and validate an automated Convolutional Neural Network (CNN) model for COVID-19 severity classification.
• To assess the effectiveness of the CNN model in categorizing patients into four severity levels: mild, moderate, severe, and critical.
• To establish a novel, automated approach for rapid COVID-19 severity assessment using chest X-ray images.

Main Methods:

• A novel Convolutional Neural Network (CNN) architecture was designed and implemented for image analysis.
• The CNN model was trained and tested on a large dataset of chest X-ray images from COVID-19 patients.
• Hyper-parameters of the CNN model were automatically tuned using a grid search optimization technique.

Main Results:

• The automated CNN model achieved an average accuracy of 95.52% in classifying COVID-19 severity.
• Experimental results demonstrated the model's effectiveness on a substantial number of chest X-ray images.
• The study successfully categorized patients into four distinct severity classes (mild, moderate, severe, critical).

Conclusions:

• The proposed CNN model offers an effective and automated solution for COVID-19 severity assessment.
• This approach can aid clinicians and radiologists in managing patient flow and treatment decisions.
• This study represents a significant advancement in automated disease severity classification using medical imaging.