An Effective Identification of Tuberculosis in Chest X-rays Using Convolutional Neural Network (CNN) Models

Abstract

Tuberculosis is a major cause of premature mortality. Tuberculosis (TB) remains a persistent global health concern, with millions of individuals affected every year. This paper provides an overview of the multifaceted nature of TB, touching on its historical significance, current global impact, diagnosis, treatment, and ongoing ef- forts to combat the disease. Diagnostic methods for TB have evolved significantly, with molecular techniques offering increased accuracy and speed. Treatment regi- mens have also seen advancements, with shorter and more effective options on the horizon. However, barriers to diagnosis and treatment, especially in underserved populations, persist. Achieving a world free of Tuberculosis is an ambitious but necessary aspiration, and ongoing efforts are crucial in making this vision a reality. The modern development of computer techniques has accelerated the TB diagnosis process. We explore the effectiveness and efficiency of sampling with cross-validation in training the network and find its outstanding effect in medical images classification.