TUBERCULOSIS DIAGNOSIS USING DEEP LEARNING

Abstract

The project aims to develop a CNN model for classifying X-ray images. In this report, we introduced a deep learning-based approach to automatically detect the manifestations oftuberculosis using chest X-rays. Pakistan is the country with the highest exposure to TB cases. Chest X-rays are used to diagnose active tuberculosis in symptomatic patients. This method ofscreening is ideally performed in primary health centres where clinicians are available and sometimes by portable X-ray machine. The main challenge ofthis screening method is timely reporting and follow-up ofthe patient at the beginning of treatment. We have created a convolutional neural network to model automated tuberculosis diagnosis, an advanced deep learning algorithm. The suggested method will automatically detect whether the given image is infected with TB or not. This method helps doctors to make accurate predictions ofthe disease in a short period oftime, thus helping to improve the clinical outcome.