A Novel Lightweight Approach to COVID-19 Diagnostics Based on Chest X-ray Images

Area of Science:

• Medical Imaging
• Machine Learning
• Computational Biology

Background:

• Developing effective diagnostic tools for COVID-19 is crucial.
• X-ray imaging offers a accessible method for disease detection.
• Machine learning can enhance the accuracy and speed of medical diagnoses.

Purpose of the Study:

• To present a novel, lightweight machine learning approach for COVID-19 diagnostics using X-ray images.
• To evaluate the effectiveness of various machine learning models in classifying COVID-19 cases from X-ray data.

Main Methods:

• Utilized real-world X-ray images from patients confirmed positive or negative by PCR tests.
• Employed a convolutional neural network for feature extraction from X-ray images.
• Applied Random Forest, XGBoost, LightGBM, and CatBoost for sample classification.

Main Results:

• The LightGBM model demonstrated superior performance in classifying COVID-19 cases.
• Achieved perfect scores: 1.00 accuracy, 1.00 precision, 1.00 recall, and 1.00 F1-score.
• Indicated high reliability of the model in differentiating COVID-19 from non-COVID-19 cases.

Conclusions:

• The proposed machine learning schema can serve as a valuable support tool for radiologists.
• The approach is computationally efficient, fast, and not overly complex.
• Potential for improving the diagnostic process for COVID-19 using medical imaging and AI.