Abstract
Background: Tuberculosis (TB) has become a global pandemic, and its eradication requires efficient screening methods, diagnostic tests, and effective drugs. Artificial intelligencebased Computer-aided Diagnostic (CADx) systems are purported to play a significant role in the mass screening of TB.
Discussion: The research on the development of CADx systems started four decades ago, and a large number of CADx systems have been developed till date. However, no independent survey focussing on the advancements in these systems has been presented. This paper fills this gap by consolidating the advancements and presents a comprehensive survey of CADx systems for TB detection developed till date with a focus on their underlying principles. It also discusses a practical model using which CADx systems can be used for screening TB in places where medical facilities and experts are not adequately available.
Conclusion: The paper also presents an overview of the current state of deep learning-based CADx systems. The development of these systems will remain in focus in the near future and will improve state-of-the-art performance in various medical domains.
Keywords: Computer-aided diagnosis, CAD systems, tuberculosis, medical imaging, chest X-rays, chest radiographs, machine learning, deep learning.
Graphical Abstract