Medical diagnosis is heavily based on machine learning (ML), especially in the field of artificial intelligence (AI). AI is distinguished by its ability to learn on its own, without the need for traditional hand-coded rules, from datasets that include text, images, and videos. Our study focuses on segmentation and categorization of pulmonary diseases (PDs) with the help of ML approaches. We extract features using the gray-level co-occurrence matrix, segment the data using Otsu-region-based thresholding, and then classify the data using artificial neural networks such as probabilistic and backpropagation neural networks. Based on chest radiographs, our suggested model shows 97.5% accuracy and a mean square error of 12.9% in diagnosing PDs, such as C19, healthy patients, lung cancer, PN, PTX, and tuberculosis.