Detection of Pulmonary Tuberculosis Using Neural Network with Feature Extraction of Gray Level Run-Length Matrix Method on Lung X-Ray Images

Abstract

Tuberculosis is a disease caused by Mycobacterium tuberculosis, which attacks the respiratory system. The purpose of the research is to detect bacteria on chest X-ray images. The x-ray images will be processed in the four stages used in this study i.e. pre-processing, segmentation, feature extraction using GLRLM (Gray Level Run Length Matrix) and detection methods using an artificial neural network. The accuracy rate for detecting tuberculosis in this research is 98.8%, with normal lungs at 97.5% and tuberculosis lungs at 100%. The input images in this study were X-ray images of the lungs in patients with positive tuberculosis and normal conditions obtained from Kaggle. The images used are 2099 images which are divided into 2019 training images, 334 tuberculosis training images, and 165 normal training images. The testing is 80 images, which consist of 40 tuberculosis images and 40 normal images. Based on the results of percentage accuracy, it can be said that the system created is very good for detecting tuberculosis using the GLRLM feature extraction method and Artificial Neural Network.