Abstract
Lung diseases, including pneumonia and COVID-19, are prevalent globally, necessitating early diagnosis for effective treatment. Medical imaging is widely regarded as an effective method for detecting lung diseases. Numerous researchers have dedicated their efforts to developing advanced detection techniques, significantly contributing to the prevention and management of these conditions. Despite advancements in imaging diagnostic methods, chest radiographs remain pivotal due to their cost-effectiveness and rapid results. This study proposes an automated system for detecting multiple lung diseases in x-ray and CT scans using a customized convolutional neural network (CNN) alongside pretrained models and an image enhancement approach. The dataset used comprises 6400 images sourced from Kaggle, categorized into three classes: pneumonia, COVID-19, and normal. To address dataset imbalance, data augmentation techniques were applied. The model includes preprocessing and classification stages, achieving high performance metrics: 96% precision, 95.33% recall, 95.66% F1-score, and 97.24% accuracy, highlighting its effectiveness compared to other deep learning models.