Lung boundary detection for chest X-ray images classification based on GLCM and probabilistic neural networks

Abstract

Extraction of various structures from the chest X-ray (CXR) images and abnormalities classification are often performed as an initial step in computer-aided diagnosis/detection (CAD) systems. The shape and size of lungs may hold clues to serious diseases such as pneumothorax, pneumoconiosis and even emphysema. With the growing number of patients, the doctors overwork and cannot counsel and take care of all their patients. Thus, radiologists need a CAD system supporting boundary CXR images detection and image classification. This paper presents our automated approach for lung boundary detection and CXR classification in conventional poster anterior chest radiographs. We extract the lung regions, sizes of regions, and shape irregularities with segmentation techniques that are used in image processing on chest radiographs. From CXR image we extract 18 features using the gray level co-occurrence matrix (GLCM). It allows us to classify the CXR image as normal or abnormal using the probabilistic neural network (PNN) classifier. The proposed method has competitive results with comparatively shorter training time and better accuracy.