Early-stage diagnosis of infectious lung diseases is crucial for increasing survival rate of patients, preventing spread of disease and decreasing diagnosis and treatment cost. Pneumonia and tuberculosis are two of the most murderous examples of lung diseases.

Pneumonia caused as a result of COVID-19, non-COVID-19 viruses, bacteria and Fungi and tuberculosis caused by Mycobacterium tuberculosis have been a major challenge for medical and healthcare sectors in many underdeveloped countries and remote communities with limited diagnosis tools and treatment approach. Molecular testing approaches based on Anti-body-based approach and PCR techniques are currently the standard techniques employed by medical expert for diagnosis of disease caused by bacteria and viruses. Other techniques include microscopy, culture test, sputum test, complete blood count. These techniques are still hindered by the utmost need of highly skilled professionals, equipped and sophisticated tools, chemical reagents and thus limit point of care diagnosis. Thus, there is need for development of fast, cheap, simple and accurate detection approach for diagnosis and predictions of these diseases.

In this project, the aim has been automatic detection pneumonia and tuberculosis. Coherence Tomography (OCT) scans and chest X-ray (CXR) images with total dataset of 5853 images of pneumonia cases caused by different strains of bacteria, non-COVID-19 viruses (such as Influenza virus, staphylococcus aureus etc.) were implemented in the project. The model is trained to classify pneumonia and tuberculosis cases. Our results are in line with the notion that CNN models can be used for classifying medical images with higher accuracy and precision. These models can now serve as a confirmation system for diagnosis of both pneumonia by and tuberculosis and offer an alternative to relieve the heavy and tedious workload experienced by radiologists and pathologists.