A neural network based computer-aided diagnosis of emphysema using CT lung images

Abstract

Chronic Obstructive Pulmonary Disease (COPD) is a disease in which the airways and tiny air sacs (alveoli) inside the lungs are partially obstructed or destroyed. The result is labored breathing. There are varying degrees of this illness, and different names for them, but it all comes back to damaged airways and air sacs. Emphysema is what occurs as more and more of the walls between air sacs get destroyed. Instead of having lots of little sacs, the sacs break up and what is left are larger sacs. These bigger sacs have less surface area for the exchange of oxygen and carbon dioxide than the tiny ones. Poor exchange of oxygen and carbon dioxide causes shortness of breath. At present, diagnosis of emphysema is done by using spirometry, X-rays, spiral chest CT-scan, bronchoscopy, pulse oximetry and arterial blood gas sampling. This paper proposes a computer-aided diagnostic system for emphysema that segments the lungs into multiple square regions and classifies the segmented regions into 5 classes of severity. The proposed algorithm is divided into three stages: 1. digital image processing, 2. feature extraction, and 3. classification using neural network (NN). The aim of this paper is to analyze the severity of the lungs region by region along with NN classification.