Abstract:
The main objective of this study is to detect tuberculosis in chest x-ray images using a convolutional neural network. Tuberculosis (TB) is an infectious disease that generally attacks the lungs and causes death for millions of people annually. Chest radiography and deep-learning-based image segmentation techniques can be utilized for TB diagnostics. Convolutional Neural Networks (CNNs) have shown advantages in medical image recognition applications as powerful models to extract informative features from images. The application of CNNs for image classification has significantly increased prediction accuracy rates. Several convolutional neural networks (CNNs) such as VGG work by building a pre-trained model that is easy to set up with minimal preprocessing. It uses libraries with weights containing millions of images to train the model before application on the actual data. This process is also called transfer learning. This thesis presents a ConvNet model that uses VGG16 for classifying CXR images, the ConvNet model is applied to the chest X-ray (CXR) dataset to identify if the patient has Tuberculosis (TB), applying such a model bypasses the requirement of building sophisticated segmentation algorithms which could be time-consuming, require professional expertise, and are mostly specialized making them inadmissible for application to other similar problems, the model can achieve accuracy of 92%. The accuracy obtained is comparable to previous work done on the dataset.
