A Frame Work For The Tuberculosis Detection Through CXR Using Deep Learning Networks

Mycobacterium tuberculosis, the causative agent of tuberculosis (TB), continues to be a serious worldwide health concern, killing about 1.3 million people in 2022, predominantly in low- and middle-income countries. The urgent need for automation in healthcare has spurred the development of supplementary tools. We have created a method for detecting TB in chest X-rays using advanced techniques such as pre-processing, picture segmentation, augmenting the data, along with Deep Learning model for classification. Our approach involves segmenting the lung region from X-rays using a U-Net model, followed by classification with a convolutional neural network (CNN). This optimized design enhances accuracy and efficiency, making it suitable for mobile deployment and providing a practical solution for TB detection in resource-limited settings.