Automated Detection of Tuberculosis and Pneumonia Using Chest XRays and Clinical Symptoms 

The Automated Detection of Tuberculosis and Pneumonia Using Chest X-rays and Clinical Symptoms aims to create an advanced diagnostic system for identifying two critical respiratory diseases—Tuberculosis (TB) and Pneumonia. This system combines chest X-ray images and clinical symptom data to provide a more accurate and reliable diagnosis. For the image data, the system uses a Convolutional Neural Network (CNN) or alternatively MobileNet, a lightweight CNN variant, to classify chest X-ray images into three categories: Normal, TB, and Pneumonia. MobileNet offers a more efficient solution for edge computing with faster inference time, making it ideal for real-time applications. For the clinical symptom data, the system employs an Artificial Neural Network (ANN) or Random Forest (RF), a robust machine learning algorithm, to analyze symptom features such as fever, cough, and fatigue for disease prediction. The Late Fusion approach is employed to combine the outputs from both models. Once the individual models make their predictions, a majority voting or secondary classifier integrates the results to provide the final diagnosis. This hybrid approach ensures that the system benefits from both visual data (X-rays) and structured symptom data, enhancing its diagnostic accuracy. The model’s ability to handle cases where one data type may be less conclusive makes it particularly effective for early detection. This integrated diagnostic system presents a promising solution for TB and Pneumonia detection, crucial for early treatment and disease management.

Keywords: Automated Detection, Tuberculosis, Pneumonia, Chest X-rays, Clinical Symptoms, Convolutional Neural Network, MobileNet, Artificial Neural Network, Random Forest, Late Fusion, Majority Voting, Hybrid Approach, Disease Prediction, Early Detection, Real-time Application, Diagnostic Accuracy.