This paper introduces a highly robust tuberculosis (TB) classification model based on the esteemed AlexNet architecture, with a primary emphasis on its exceptional performance in TB detection. Our model exhibits significant promise for early TB diagnosis, especially within resource-constrained healthcare environments. Employing subtle architectural enhancements and rigorous training on a diverse dataset encompassing a spectrum of TB symptoms, our AlexNet-based model adeptly segregates medical images into two pivotal categories: “normal” and “tuberculosis-infected.” With a classification accuracy rate of 99.67%, the achieved results underscore the model's robustness, substantiating its potential to advance TB diagnostics. This contribution ultimately aims to enhance patient care and further global health initiatives. By addressing the critical challenge of early TB detection, our research not only advances the field of medical image analysis but also aligns with global health initiatives aimed at reducing the burden of tuberculosis worldwide. The combination of accuracy, efficiency, and detailed analysis offered by our model has the potential to revolutionize TB diagnostics, ultimately improving patient outcomes and contributing to the overarching goal of eradicating TB as a public health threat. As we continue to refine and optimize our TB classification model, we remain dedicated to the pursuit of enhanced patient care and the advancement of global health efforts.