CovidSORT: Detection of Novel COVID-19 in Chest X-ray Images by Leveraging Deep Transfer Learning Models

Tammina, Srikanth (2022) CovidSORT: Detection of Novel COVID-19 in Chest X-ray Images by Leveraging Deep Transfer Learning Models. In: 2nd International Conference on Data Science, Machine Learning and Applications, ICDSMLA 2020, 21 November 2020 through 22 November 2020, Pune.

Full text not available from this repository. (Request a copy)


Coronaviruses are a cluster of viruses belonging to the family of Coronaviridae, which infect animals and humans. Coronaviruses related to humans can cause mild disease very similar to common flu, which others cause more severe acute diseases. The new COVID-19, which was first detected in the city of Wuhan in Hubei province, China in December 2019. This new coronavirus that previously has not been identified emerged in China when a cluster of Pneumonia cases was reported. Signs and symptoms include respiratory symptoms namely, chronic mucus, fever, lingering chest pain, stubborn cough, breathing noisily. In more severe cases, the virus can lead to cause pneumonia, acute respiratory distress syndrome (ARDS), and sometimes death. The Health care system and the global economy has been severely disrupted since the Covid-19 pandemic began. An early diagnosis to identify the infection is very crucial to mitigate the stress on the health care system and health care providers. A chest X-ray is performed on patients to detect any inflammation in the lungs of a human. The objective of this paper is to leverage artificial intelligence models coupled with image augmentation techniques to accurately classify the chest X-ray images into two classes namely, Pneumonia and Normal. In this research, a new framework, CovidSORT, is proposed for detecting pneumonia infected lungs using chest X-ray images. The proposed framework is developed using deep transfer learning models namely, Inception-V3, VGG16, VGG19, ResNet-50, DenseNet-121 and MobileNetV2 which were pre-trained on ImageNet which led to quicker model training. Additionally, these models are fine-tuned with image augmentation techniques for better accuracy. The research concludes that the ensemble model built on the majority voting approach from these models in identifying pneumonia has achieved a classification accuracy of 96.83%. The above framework can be used by radiologists to corroborate and identify COVID patients quickly. © 2022, The Author(s), under exclusive license to Springer Nature Singapore Pte Ltd.

[error in script]
IITH Creators:
IITH CreatorsORCiD
Item Type: Conference or Workshop Item (Paper)
Uncontrolled Keywords: Chest X-ray images; Convolutional neural network; Deep learning; Ensemble; Image augmentation; Image classification; Inception-v3; Machine learning; Transfer learning
Subjects: Electrical Engineering
Divisions: Department of Electrical Engineering
Depositing User: . LibTrainee 2021
Date Deposited: 28 Jul 2022 12:01
Last Modified: 28 Jul 2022 12:01
Publisher URL:
Related URLs:

    Actions (login required)

    View Item View Item
    Statistics for RAIITH ePrint 9994 Statistics for this ePrint Item