Lung Disease Prediction using Deep Learning Techniques

Abstract

The timely and correct diagnosis of lung ailments has been an important challenge in the medical domain because of the similarity in symptoms and the inability to analyze chest X-ray images manually. However, the concept of automated image analysis in the medical domain by utilizing deep learning techniques has opened new doors for improving diagnosis accuracy and efficiency. In this paper, a new automated system has been introduced utilizing the concept of Convolutional Neural Networks for multi-class classification of lungs into bacterial pneumonia, Viral pneumonia, Tuberculosis, COVID-19, and Normal lungs. Advanced models of neural networks such as DenseNet201, ResNet50, VGG16, and Inception V3 are used in this paper for feature extraction from chest X-ray images for providing accurate classification. The models of AI and machine learning in this paper prove to have remarkable ability for distinguishing between different types of lung ailments without relying heavily on manual analysis by experts in the domain of medicine.