Detection and classification of pneumonia in chest X-ray images by supervised learning
Shahida Parveen, Khan Bahadar Khan
Abstract
Pneumonia known as a “silent killer” is a worldwide health issue, causing a large number of mortality. In developing countries, the indication of this lung's disease makes it one of the deadliest among children under the age of five years old. The numerical contrast between infection rates and death rates shows how much the early diagnosis of pneumonia is crucial. To overcome this crucial issue, accurately and fast diagnose is required for timely treatment. The Chest X-rays (CXR) is the best choice in the term of cost and availability to diagnose pneumonia. This paper presents a supervised computer aided diagnostic (CAD) system for the classification of the infected lung with pneumonia and the normal X-ray image. The CAD system can process hundreds of X-rays images to extract features using the Histogram of Oriented Gradient (HOG) technique. Then, we trained three different classifier: Support Vector Machine (SVM), decision Tree and Random Forest to classify the extracted features. The largest publicly available chest X-ray dataset contains 5,232 frontal view images is used for testing and validation. Performance parameters accuracy, recall, precision and F1-score are compared with the existing literature to check the effectiveness of the proposed model.