Enhanced Tuberculosis Detection Using Deep Learning and GAN Data Augmentation Techniques for Chest x-Ray Analysis

With around 1.6 million deaths yearly, primarily in low-resource environments, abstract-tuberculosis (TB) is a serious worldwide health issue. While human analysis can be error-prone and labor-intensive, early discovery using chest X-rays is vital. We used advanced machine learning methods including a mix of Histogram of Oriented Gradients (HOG) and Convolutional Neural Networks (CNN), alongside Deep Convolutional Generative Adversarial Networks (DCGANs) to create synthetic TB images, so helping to balance the dataset and reduce class imbalance. We applied explainable artificial intelligence (XAI) methods including LIME and Grad-CAM for greater interpretability in order to improve model transparency. The outcomes showed notable progress; our hybrid model obtained a stunning 99.5% diagnosis accuracy. With respective accuracy of 96.87% and 94.67%, models such as Xception and MobileNetV2 also fared rather well. Especially in resource-limited healthcare settings, this study provides a consistent and interpretable strategy for tuberculosis screening.