Title: Analysis of chest X-ray images using deep learning approaches

Authors: Ruchika Arora; Indu Saini; Neetu Sood

Addresses: Department of ECE, Dr. B R Ambedkar National Institute of Technology, Jalandhar, 144011, India ' Department of ECE, Dr. B R Ambedkar National Institute of Technology, Jalandhar, 144011, India ' Department of ECE, Dr. B R Ambedkar National Institute of Technology, Jalandhar, 144011, India

Abstract: Common thorax diseases such as pneumonia, tuberculosis, are diagnosed with digital radiography, i.e., chest X-ray (CXR) images. This paper provides a glimpse of chest abnormalities classification and annotation methods for CXR images that improves work efficiency and diagnosis accuracy. At present, pre-trained models such as ResNet, DenseNet, and its variants have become important deep learning (DL) approaches for successful classification and detection of diseases. This detailed literature review highlights need for integration of both image and text metadata features for designing multi-label image classification systems for effective diagnosis of chest diseases. As novel coronavirus disease (COVID-19) causes lung problems, so a new research frontier is to fight against COVID-19. This paper covers an insight into literature review of DL algorithms used for diagnosis of COVID-19, and also emphasises move from computer aided detection (CAD) to the clinic illustrating recent practices, problems, and up-to-date information on CXR image classification and annotation.

Keywords: chest X-ray images; deep learning; machine learning; ResNet; DenseNet; image classification; disease detection; image captioning; COVID-19.

DOI: 10.1504/IJMEI.2023.132576

International Journal of Medical Engineering and Informatics, 2023 Vol.15 No.4, pp.311 - 322

Received: 30 Dec 2020
Accepted: 25 May 2021
Published online: 30 Jul 2023 *

Full-text access for editors Full-text access for subscribers Purchase this article Comment on this article