A Survey on Lung Disease Prediction Using a Hybrid Capsule Network-VGG19 Model with Deep Learning

Abstract

Lung diseases affect the respiratory system, causing issues with breathing and oxygen exchange. Common conditions include lung cancer, tuberculosis, COVID-19, and pneumonia, each affecting lung tissues in different ways. These diseases can lead to symptoms like coughing, shortness of breath, and chest pain, and often require medical imaging like chest X-rays and CT scans for diagnosis and monitoring. A hybrid model is proposed that combines a Capsule Neural Network with the VGG19 architecture to improve the classification of lung diseases. The Capsule Network is designed to capture spatial hierarchies in images, significantly reducing the risk of misclassification by understanding the relationships between different features. On the other hand, VGG19 enhances feature extraction, providing a deeper analysis of the images. This innovative approach not only improves prediction accuracy but also simplifies the classification process, enabling the inclusion of additional disease classes. By leveraging the strengths of both Capsule Networks and VGG19, the proposed AI-based system aims to classify lung diseases with higher reliability and accuracy. It enhances early-stage diagnosis by efficiently capturing image relationships and supporting multi-disease detection, ultimately overcoming the limitations of existing methods in the field of medical imaging.