Tuberculosis (TB) remains one of the leading infectious diseases worldwide, making early and accurate diagnosis essential for effective treatment. This study presents a comprehensive data-driven approach for TB detection using chest X-ray imaging combined with optimized deep learning models. The proposed framework integrates advanced preprocessing, feature extraction, and model optimization techniques to improve classification accuracy and robustness. A curated dataset of chest radiographs is used to train and validate the deep learning architecture, ensuring reliable detection performance across diverse patient groups. Experimental results demonstrate that the optimized model significantly enhances diagnostic precision compared to conventional methods. This approach provides a scalable, efficient solution that supports clinical decision-making and enables automated TB screening programs.