Detection of Pneumonia from X-ray Images using Eigen Decomposition and Machine Learning techniques

The mortality rate due to pneumonia is very high across the world and particularly in India. This respiratory and lung-based disease can be controlled by early detection and treatment. Many reported articles are available on this topic. In the current investigation an attempt has been made to develop and test k-nearest neighbor (k-NN), random forest (RF), and light gradient boosting machine (LGBM) based pneumonia detection models using novel eigen domain extracted features from the X-ray images of patients. By conducting simulation-based experiments on the developed models using publicly available standard data set, the various performance matrices have been obtained and compared. The analysis of results demonstrates that the proposed LGBM with eigen domain features provide the best classification performance compared to other reported models.