Topics in Artificial Intelligence Applied to Industry 4.0

4Detection from Chest X‐Ray Images Based on Modified Deep Learning Approach

Jyoti Dabass¹, Manju Dabass², and Ananda K. Behera³

¹ DBT Centre of Excellence Biopharmaceutical Technology, IIT, New Delhi, India

² EECE Department, The Northcap University, Gurugram, Haryana, India

³ Artificial Intelligence and Machine Learning Programme, Liverpool John Moores University, Liverpool, UK

4.1 Introduction

A lung condition known as “tuberculosis” (TB) is caused by the Mycobacterium tuberculosis bacteria. After COVID‐19, it is the second most contagious illness that spreads through airborne droplets from a TB patient's coughing and sneezing. The World Health Organization (WHO) guesstimates that 10 million people would be infected by TB by 2020, and 1.5 million will die from it. TB is a curable disease and can be treated through four antimicrobial drugs taken over six months [1]. To avoid the spread of the infection and treatment of TB‐affected patients, timely diagnosis and detection of TB are necessary. Currently used septum smear microscopy and WHO‐recommended rapid molecular assay tests to detect TB are both costly and time taking [2]. The expense of these tests restricts the detection of the disease because the majority of TB‐afflicted persons live in low‐ to middle‐income South Asian nations. Expert radiologists may additionally detect TB with the use of chest X‐ray (CXR) images. However, the procedure of diagnosing TB in such regions is hampered by the shortage of skilled radiologists. ...