Diagnosis of COVID-19 Cases from Chest X-ray Images Using Deep Neural Network and LightGBM

TL;DR

This paper presents an automated COVID-19 diagnosis method from chest X-ray images using a deep neural network for feature extraction, feature selection, and LightGBM for classification, achieving high accuracy on a public dataset.

Contribution

The study introduces a novel combination of DensNet169, ANOVA feature selection, and LightGBM for COVID-19 detection from X-ray images, demonstrating improved accuracy.

Findings

Achieved 99.20% accuracy in two-class classification.

Achieved 94.22% accuracy in multi-class classification.

Validated on the ChestX-ray8 dataset.

Abstract

The Coronavirus was detected in Wuhan, China in late 2019 and then led to a pandemic with a rapid worldwide outbreak. The number of infected people has been swiftly increasing since then. Therefore, in this study, an attempt was made to propose a new and efficient method for automatic diagnosis of Corona disease from X-ray images using Deep Neural Networks (DNNs). In the proposed method, the DensNet169 was used to extract the features of the patients' Chest X-Ray (CXR) images. The extracted features were given to a feature selection algorithm (i.e., ANOVA) to select a number of them. Finally, the selected features were classified by LightGBM algorithm. The proposed approach was evaluated on the ChestX-ray8 dataset and reached 99.20% and 94.22% accuracies in the two-class (i.e., COVID-19 and No-findings) and multi-class (i.e., COVID-19, Pneumonia, and No-findings) classification problems, respectively.