COMPARATIVE ANALYSIS OF SOLITARY LUNG NODULE CLASSIFICATION USING DIFFERENT FUNCTIONS OF BACK PROPAGATION NEURAL NETWORK

One of the deadliest diseases across the global human population is lung cancer. Computed Tomography (CT) is preferred over X- ray detection for lung cancer to obtain improved accuracy. In this paper, a comparison is made among different training functions of Back Propagation Neural Network (BPNN) for classifying the solitary lung nodule as normal and abnormal. Gaussian filter is used for image preprocessing. The lung parenchyma region is identified using Active Contour Model (ACM). The feature extraction process uses Gray-level Co-occurrence Matrix (GLCM) method. The BPNN classifier confirms whether the nodule is normal or abnormal.