Tuberculosis (TB) is an infectious disease usually caused by Mycobacterium tuberculosis (MTB) bacteria. Tuberculosis generally affects the lungs, but can also affect other parts of the body. Most infections show no symptoms, in which case it is known as latent tuberculosis.
In Low and Middle-Income Countries (LMICs), efforts to eliminate the Tuberculosis (TB) epidemic are challenged by the persistent social inequalities in health, the limited num- ber of local healthcare professionals, and the weak healthcare infrastructure found in resource-poor settings. The modern development of computer techniques has accelerated the TB diagnosis process.
We design a novel method to apply CNN models to detect and classify TB manifestations in X-ray images. Based on the research result and the specific technical problems in this large unbalanced, less-category dataset, we use a set of optimization solutions to further improve the accuracy. Our method improves the accuracy for classifying multiple TB manifestations by a large margin. Our methods and results show a promising path for more accurate and faster TB diagnosis in LMICs healthcare facilities.