Lung Tuberculosis Detection Using X-Ray Images

This project work is based on several experiments for lung tuberculosis detection that were conducted implementing Ensemble, Inception, DenseNet, and classification algorithms. The most lethal infectious disease in the world is lung tuberculosis, a bacterial infection. All around the world, there are two billion tuberculosis patients. Mycobacterium tuberculosis, also referred to as Tubercle bacillus, is the bacterium that causes lung tuberculosis. This project effort looks for ways to help patients save a lot of money when they need an additional opinion for a result that has already been determined. Once we have X-ray images, we train and evaluate them before uploading them to classification algorithms. These filters assist in obtaining delicate textural details and in removing unnecessary noise. In order to identify lung tuberculosis, we used KNN, Modified SVM, Decision Tree, and Naive Bayes classification algorithms. According to the project’s results, the modified SVM classifier performs better than the others.