A Review of Deep Learning-Based Detection Methods for Tuberculosis

Tuberculosis (TB) causes exceptionally high mortality rates, and early identification of TB is the key to saving patients. Deep learning techniques have proven to be an essential tool to assist radiologists in detecting abnormalities and multiple diseases. This study categorizes and analyzes deep learning-based techniques for TB diagnosis. Available public datasets are presented, and each method’s performance is compared comprehensively for the use of future researchers. Finally, we explore the challenges of detecting TB using deep learning algorithms and the future research prospects in this field.