Abstract: This text's research investigates the automation of processes that diagnose tuberculosis and pneumonia from medical imaging data using convolutional neural networks (CNNs). Worldwide, TB and pneumonia represent serious health risks that need to be identified quickly to be effectively treated and controlled. We introduce a CNN-based approach that leverages deep learning to increase the efficiency and accuracy of diagnosis. We use publicly available datasets of chest X-ray images to train and evaluate the CNN model. Based on meticulous testing and analysis, our findings indicate that this is the most reliable method for diagnosing tuberculosis or pneumonia patients. The CNN model outperforms conventional approaches for real-time clinical applications because of its superior performance metrics, which include high accuracy, sensitivity rates, and specificity values. This research contributes to ongoing attempts to use artificial intelligence for improved medical diagnosis, enabling early illness recognition and better patient outcomes for the treatment of both tuberculosis and pneumonia.

Keywords: Convolutional Neural Networks (CNN), Tuberculosis, Pneumonia Patients, Meticulous Testing, Sensitivity Rates

| DOI: 10.17148/IMRJR.2025.020301