Lung cancer classification using radial basis function neural network model with point operation

In this paper, we propose radial basis function neural network (RBFNN) model to classify the lung cancer through lung photo/image called chest X-ray. We also present the benefit of image improvement of point operation. The point operation specifically aims to improve the intensity of the lung image. The RBFNN model involves two kind of learning processes. Here, we consider K-means clustering and global ridge regression methods to learn the center and the width parameters and also the weights of RBFNN model, respectively. Three classifications of lung conditions including normal, benign, and malignant are examined. The structure of RBFNN model is generated based on the parameters resulted from the images extraction Gray Level Co-occurrence Matrix (GLCM) method. The experimental result shows the superiority of RBFNN model with point operation over the RBFNN model without point operation.