Feature Fusion Based Ensemble Method For Multi-Level Lung Diseases Classification Of X-Ray Images Using Deep Learning

Abstract

In recent years, infectious respiratory illnesses have surpassed all other causes of death in the world. Pneumonia, Tuberculosis (TB) and COVID-19 are the most severe and prevalent infectious respiratory disorders caused by bacteria and virus that typically affect the lungs which can even lead to death. Currently, COVID-19 is ranked as the highest cause of death in recent years as the mortality rate crosses 6 million. The most interesting and complicating fact about these respiratory diseases are the similarity in their symptoms. So, it is necessary to classify all the three diseases which can be accomplished by applying deep learning techniques in Chest X Ray (CXR) images of patients. Deep learning based multi-level classification is carried out in this research to detect whether the patient is affected by Pneumonia or Tuberculosis or COVID-19. Features plays a major role in classification and so it is decided to develop a novel feature extraction technique namely “Fusion of Handcrafted and Deep features” (FHD). The proposed FHD technique generates a new ensemble Feature Vector (FV) by concatenating handcrafted and deep features and it achieves a classification accuracy of 96.2% using ensemble FV. Handcrafted features include texture features obtained from Gray-level co-occurrence matrix (GLCM), Grey Level Difference Matrix (GLDM) and Gray Level Size Zone Matrix (GLSZM). Extraction of deep features and classification is done by Modified XceptionNet. The results obtained using the proposed technique is reliable and effective, hence radiologists can utilize this method to detect lung disorders using CXR images.