Multi-Class Classification of Lung Disease Using X-ray Images

Abstract

Chest x-ray are commonly used medical imaging technique for medical diagnosis. For lung disease many people died every year. In this crisis, an automated system is needed for detect lung related illness. An automated system helps to reduce reading errors, quick report delivery and decrease work pressure. In this research, seven pre-trained model was applied on a merged dataset and showed these comparisons. In this dataset these are four class which are Covid19, Pneumonia and Normal class. After the pre-processing steps x-ray images were fed for classification in VGG16, VGG19, Xception, InceptionV3, DenseNet121, MobileNet and RseNet101. VGG16 achieved the highest accuracy which was 95%. Keywords: X-ray images; Computer aided diagnosis; lung disease; vgg16; deep learning