A Systematic Literature Review of Artificial Intelligence on Medical Imaging: COVID-19 and Tuberculosis Classification

Medical imaging techniques play a pivotal role in disease management and monitoring. Image scans offer a magnified lens into the intricate workings of human body parts with clear, precise information, and fast image acqui-sition. In particular, chest imaging reveals lung conditions, including COVID-19 and tuberculosis. However, even skilled radiologists may find it challenging to evaluate minor variations in the amount and nature of lung abnormalities. Artificial intelligence (AI) emerges as a promising solution. AI can support conventional medical imaging equipment by providing computational power to process images more quickly and accurately. Despite this potential, comprehensive studies on AI's benefits in medical imaging remain scarce, especially for COVID-19 and tuberculosis. These conditions share structural similarities in their radiological patterns, emphasizing the need for targeted research. To address this research gap, this review paper provides an AI-powered method of tracking, diagnosis, and prognosis of COVID-19 and tuberculosis using different types of medical imaging scans. Several models, including deep learning architectures and convolutional neural network (CNN), are examined in this comprehensive review. The analysis demon-strates how well they classify lung ailments; certain models have accuracy rates as high as 98.80 % accuracy for TB and 98.31% for COVID-19. However, there are still issues, namely the improvement of AI transparency and its incorporation into clinical practice. Reducing diagnostic errors and enabling faster treatment are two ways in which addressing these concerns could improve patient care.