COVID-19 Severity Classification on Chest X-ray Images

TL;DR

This study develops a machine learning approach to classify COVID-19 severity levels from chest X-ray images, utilizing image preprocessing, data augmentation, and deep learning models, achieving high accuracy and reliability.

Contribution

It introduces a novel method for classifying COVID-19 severity using pre-trained deep learning models and data augmentation, filling a gap in existing diagnostic AI tools.

Findings

ResNet-50 achieved 95% accuracy

High recall and F1-score indicate reliable classification

Data augmentation improved model performance

Abstract

Biomedical imaging analysis combined with artificial intelligence (AI) methods has proven to be quite valuable in order to diagnose COVID-19. So far, various classification models have been used for diagnosing COVID-19. However, classification of patients based on their severity level is not yet analyzed. In this work, we classify covid images based on the severity of the infection. First, we pre-process the X-ray images using a median filter and histogram equalization. Enhanced X-ray images are then augmented using SMOTE technique for achieving a balanced dataset. Pre-trained Resnet50, VGG16 model and SVM classifier are then used for feature extraction and classification. The result of the classification model confirms that compared with the alternatives, with chest X-Ray images, the ResNet-50 model produced remarkable classification results in terms of accuracy (95%), recall (0.94), and F1-Score (0.92), and precision (0.91).