Tuberculosis detection, an effective approach using convolutional neural network

In most regions of the world, tuberculosis (TB) poses a major health risk and is spread through the air. A majority of diagnostic techniques takes a long time, are not always accurate, and were mostly created in the last century. The most popular technique for early tuberculosis detection in big populations is chest X-rays. The radiologist’s experience and ability to interpret images will determine whether or not this procedure is successful. Because convolutional neural networks (CNNs) can learn both medium-level and high-level picture representations, they have become a popular deep learning technique. Several CNN models were employed in this work as Google Net models to categorize chest radiographs into TB-positive and TB-negative groups. A comparison of deep learning methods that can be used to process chest X-rays and identify tuberculosis is presented in this article. A publicly accessible dataset called the Tuberculosis (TB) Chest X ray Database is used to gauge the system’s performance.