Automated Methods for Tuberculosis Detection/Diagnosis: A Literature Review

Tuberculosis (TB) is one of the leading infectious causes death worldwide. The effective management and public health control this disease depends on early detection careful treatment monitoring. For many years, microscopy-based analysis sputum smears has been most common method to detect quantify Mycobacterium tuberculosis (Mtb) bacteria. Nonetheless, form a challenging procedure since examination can only be reliably performed by trained personnel with rigorous quality systems in place. Additionally, it affected subjective judgement. Furthermore, although fluorescence-based sample staining methods have made easier recent microscopic time-consuming operation. Over past two decades, attempts automate practice. Most approaches focused establishing an automated diagnosis, while others centred measuring bacterial load or detecting localising Mtb cells for further research phenotypic characteristics their morphology. literature incorporated machine learning (ML) computer vision as part methodology achieve these goals. In review, we first gathered publicly available TB smear microscopy image sets analysed disparities datasets. Thereafter, evaluation metrics used assess efficacy each its particular field. Finally, generated comprehensive summaries prior work ML deep (DL) detection, including review limitations.