This project focuses on the detection and classification of various lung diseases using chest X-ray images. The diseases considered in this study include Normal, Bacterial Pneumonia, Viral Pneumonia, COVID-19, Tuberculosis, and Emphysema. The dataset used for this project is the Chest X-Ray 6 Classes Dataset available on Kaggle, which contains labeled X-ray images corresponding to these six categories.

The main objective of this project is to develop an automated deep learning system that can accurately classify chest X-ray images into the correct disease category. Early and accurate detection of lung diseases is essential for effective treatment and reducing mortality rates.

To achieve this goal, several advanced deep learning architectures are implemented, including InceptionV3, CNN with Attention Mechanism, ResNet50 with Convolutional Block Attention Module (CBAM), and a Hybrid Swin Transformer with Vision Transformer (ViT) model. These models are capable of automatically extracting complex features from medical images and focusing on important regions of the lungs.

The models are trained and evaluated using standard evaluation metrics such as accuracy, precision, recall, and F1-score to measure their performance. By combining convolutional neural networks with attention mechanisms and transformer-based architectures, the proposed system aims to provide a highly accurate and efficient solution for automated lung disease diagnosis.

The developed model can assist healthcare professionals and radiologists in making faster and more reliable diagnoses, thereby improving medical decision-making and patient care.

Keywords: Chest X-ray, Pneumonia, COVID-19, Tuberculosis, Emphysema, Deep Learning, InceptionV3, ResNet50, CBAM, Vision Transformer, Swin Transformer.