No abstract

Tuberculosis (TB) is a common infectious disease caused by bacteria named mycobacterium tuberculosis, which is preventable and curable if detected early. In feature extraction of medical images, any unwanted features extracted may lead to efficiency loss. To overcome this, the features are optimized using Orthogonal Learning Particle Swarm Optimization (OLPSO) technique, which is used to identify the specific set of features from the image and ranks the features based on decision task equation. Based on which the images are classified. In addition, this paper proposes a hybrid classification to differentiate the images as Cavitary TB and Miliary TB by nomination method of classification. The hybrid classifier is an integration of Support Vector Machine (SVM) and Artificial Neural Network (ANN) which are applied to CT scan lung images to provide results with high accuracy. This experiment results show that, it is possible to identify and classify TB images by using MATLAB classifiers.