Lung cancer is a disease with a huge prevalence in a few nations around the globe. Improving the stage of early analysis and the detection of small Lung Nodule (LN) has been dependably a vital subject for imaging technologies. In this paper, the LN detection from the ELCAP public lung image database is investigated by an innovative classifier that is Radial Basis Function-Neural Network (RBF-NN). Using ROI extraction, the two sides of the lung portion is separated and then image contrast level is enhanced by the Adaptive Histogram Equalization (AHE). After that, nodules are segmented Fuzzy C-Means (FCM) with centroid selection. The proposed classifier technique is applied over segmented images; where the images are classified as nodule detected and normal lung image based on extracted feature sets. The paper improved the classification accuracy by the optimal selection of hidden nodes in RBF-NN using Enhanced Cuckoo Search Optimization (ECSO) algorithm. The performance of the proposed RBF-NN+ECSO is examined by some measures like sensitivity, specificity, and accuracy. The implementation results reveal that the RBF-NN+ECSO algorithm achieves high accuracy and less computation time compared to existing neural network algorithms.
Volume 11 | 07-Special Issue
Pages: 1354-1363