Automated Detection of Lung Tuberculosis Based on X-ray Image Analysis

Abstract

Detection of lung tuberculosis is mostly based on X-ray images. Image segmentation is important in different fields of image processing. Image segmentation is process of dividing images according to its characteristics. Different methods are presented for image segmentation. In this paper to find nodules, symptoms of diseases in X-ray images, we use watershed segmentation approach. First take the gray scale image and then applying the watershed segmentation approach to segment the image with catchment basins. When the lung image is isolated from X-ray image, the suspected nodule pixels in the lung can be found. Based on nodules and user’s input symptoms, patient can be defined either suffer from lung TB or not.