Reliable Tuberculosis Detection Using Chest X-Ray With Deep Learning, Segmentation and Visualization

TLDR

Tuberculosis is a chronic lung disease and one of the top ten causes of death worldwide, making early and accurate detection essential. This study aims to reliably detect TB from chest X‑ray images by applying image pre‑processing, data augmentation, segmentation, and deep‑learning classification. The authors assembled 7,000 X‑ray images, trained nine pretrained CNNs (including ResNet, ChexNet, DenseNet, etc.) on whole and U‑net‑segmented lung images, and visualized that the networks focus on lung regions. ChexNet achieved 96.5 % accuracy on whole images, while DenseNet201 reached 98.6 % accuracy on segmented lungs, demonstrating state‑of‑the‑art performance for rapid TB diagnosis.

Abstract

Tuberculosis (TB) is a chronic lung disease that occurs due to bacterial infection and is one of the top 10 leading causes of death. Accurate and early detection of TB is very important, otherwise, it could be life-threatening. In this work, we have detected TB reliably from the chest X-ray images using image pre-processing, data augmentation, image segmentation, and deep-learning classification techniques. Several public databases were used to create a database of 3500 TB infected and 3500 normal chest X-ray images for this study. Nine different deep CNNs (ResNet18, ResNet50, ResNet101, ChexNet, InceptionV3, Vgg19, DenseNet201, SqueezeNet, and MobileNet) were used for transfer learning from their pre-trained initial weights and were trained, validated and tested for classifying TB and non-TB normal cases. Three different experiments were carried out in this work: segmentation of X-ray images using two different U-net models, classification using X-ray images and that using segmented lung images. The accuracy, precision, sensitivity, F1-score and specificity of best performing model, ChexNet in the detection of tuberculosis using X-ray images were 96.47%, 96.62%, 96.47%, 96.47%, and 96.51% respectively. However, classification using segmented lung images outperformed that with whole X-ray images; the accuracy, precision, sensitivity, F1-score and specificity of DenseNet201 were 98.6%, 98.57%, 98.56%, 98.56%, and 98.54% respectively for the segmented lung images. The paper also used a visualization technique to confirm that CNN learns dominantly from the segmented lung regions that resulted in higher detection accuracy. The proposed method with state-of-the-art performance can be useful in the computer-aided faster diagnosis of tuberculosis.