Enhancing Feature Extraction for Lung Disease Classification and Segmentation from X Ray Images Using Pulmonnet Modules

Abstract

Lung diseases significantly impact global health, with potential causes including newlinesmoking, genetic predisposition, and pathogens. If left untreated, chronic lung diseases can newlineescalate to severe conditions like lung carcinoma or respiratory failure. Prompt screening for newlinerespiratory ailments is thus crucial as it allows for therapeutic approaches to limit the newlinecontamination of the infection, potentially reducing the mortality rate. While X-ray imaging newlineremains the primary diagnostic tool for lung abnormalities, its limitations in capturing subtle newlinedetails necessitate advanced solutions. Deep learning models have emerged as powerful newlinetools, capable of extracting intricate features from chest X-rays and offering precise and newlinerapid detection and segmentation of lung diseases. newlineThis thesis explores the integration of X-rays with AI-driven systems to automate newlinethe identification of pulmonary disorders with the objective of enhancing both accuracy and newlineprecision in medical diagnostics. This study leverages two renowned benchmark datasets newlinenamely the COVID-19 Radiography Database and the Tuberculosis (TB) Chest X-ray newlineDatabase, both sourced from the Kaggle repository to develop advanced models newline