Automated Framework for the Tuberculosis Detection and Classification in X-Ray Images Using Deep Learning Algorithm
C. Heltin Genitha, I. Kalaivani, A.S. Hepsi Ajibah, S. Jalagandeswaran, Karthigha Balamurugan
Abstract
This study focuses on utilizing deep learning algorithms to detect Tuberculosis (TB) through the examination of chest radiograph (X-ray) images. The primary goal of this research study is to develop and demonstrate a Convolutional Neural Network (CNN) based method specifically designed for the TB detection in chest images. The algorithm developed comprises of three key stages: pre-processing, feature extraction, and classification. During the initial pre-processing stage, the input chest X-ray image undergoes normalization and resizing operations to establish a standardized and consistent fixed size. This step aims to enhance the uniformity and comparability of the images. The CNN technique is utilized to automatically identify desirable characteristics from the X-ray of chest image during the feature extraction step. The chest X-ray image is classified as tuberculosis positive or negative during the classification step using the retrieved features. The findings of this study could enhance the precision and speed of tuberculosis detection utilising chest X-ray imaging. The proposed algorithm has the potential to assist healthcare professionals in the early and accurate detection of TB.