Multimodal Deep Learning for Precise Lung Pathology Discrimination

Tuberculosis (TB) is a serious airborne infectious disease with a high fatality rate, primarily affecting the lungs. Early detection is vital for effective treatment, yet the current healthcare landscape suffers from a significant shortage of medical professionals and screening facilities, which hinders timely diagnosis. The lung cancer pandemic has further exacerbated these challenges, straining healthcare systems and complicating diagnostic procedures for respiratory diseases. Existing diagnostic systems predominantly rely on the ResNet algorithm to predict TB. However, these systems often fall short when it comes to accurately diagnosing COVID- 19, leading to reduced prediction accuracy and missed opportunities for early intervention. In light of these limitations, this study proposes a novel hybrid approach that combines Recurrent Neural Networks (RNN) and Convolutional Neural Networks (CNN) to facilitate the simultaneous prediction of both TB and LUNG CANCER. The integration of RNNs allows for effective temporal analysis of sequential data, while CNNs excel in extracting spatial features from medical imaging, such as X-rays or CT scans. This synergistic model is designed to improve the predictive accuracy for both diseases by leveraging the strengths of each architecture. By enhancing feature extraction and analysis capabilities, this innovative approach aims to address the pressing need for reliable diagnostic tools in the face of inadequate healthcare resources. Ultimately, the future system has the efficiency to enable earlier diagnoses and the treatment outcomes, and contribute to better public health management during dual outbreaks of TB and lung cancer.