USING PRE-TRAINED CONVOLUTIONAL NEURAL NETWORKS, AN ANALYSIS OF COVID-19 AND PNEUMOMIA



Authors

  • T.Sundaravadivel

DOI:

https://doi.org/10.15282/jmes.17.1.2023.10.0759


Keywords:

Chest X-rays (CXR), Inception-v3, Resnet50, Random Forest, Computed Tomography (CT)


Abstract

Global crises brought the new corona virus illness outburst that started in December 2019. As of October 10, 2020, more than 200 nations have been affected by the pandemic, which has resulted in approximately 37 million confirmed infections and over a million deaths..The necessity for establishing a quick, inexpensive, and accurate screening approach is increasing as the number of cases rises. For this, Kaggle is used to collect the chest x-ray images. Inception-v3, Resnet50, two convolutional neural network architectures, are employed as feature extractors in this work. Then, the derived features are given to Random Forest. Random Forest classifies into COVID 19, Pneunomia and normal image. The features obtained from Inceptionv3 with Random Forest provides the satisfactory results of 91.11%.



Published

2024-04-08

How to Cite