Classifying Lung Cancer as Benign and Malignant Nodule Using ANN of Back-Propagation Algorithm and GLCM Feature Extraction on Chest X-Ray Images

• Published in: Wireless personal communications Vol. 126; no. 1; pp. 167 - 195
• Main Authors: Napoleon, D., Kalaiarasi, I.
• Format: Journal Article
• Language: English
• Published: New York Springer US 01.09.2022; Springer Nature B.V
• Subjects: Algorithms; Artificial neural networks; Back propagation networks; Brain cancer; CAD; Chest; Classification; Communications Engineering; Computer aided design; Computer Communication Networks; Discriminant analysis; Engineering; Feature extraction; Image classification; Image processing; Image segmentation; Lung cancer; Medical imaging; Networks; Neural networks; Patients; Performance evaluation; Signal,Image and Speech Processing; Support vector machines; Training; Chest X-ray imaging; K-means hasil image segmentation; Artificial neural network; Grey level co-occurrence matrix
• ISSN: 0929-6212; 1572-834X
• DOI: 10.1007/s11277-022-09594-1

Lung cancer is the most suffering disease which is very difficult to identify in advance and it is not easily cure if the stage of cancer becomes more malignant, the lung cancer is similar like other cancers such as breast cancer, colorectal cancer, brain tumour etc. Now-a-days, there are lot of technologies are developed to predict and treating the diseases, but still have some trouble in detecting the cancer nodule more accurately. Due to increasing in number of patients admitted in clinic, hospitals, etc., doctors cannot able to monitor every patient with high care and they failed to guide their patients with greater attention. Accordingly, the radiologists require a technology named Computer Aided Design (CAD) system for precise recognition and classification of lung nodule where the detected node is cancerous or non-cancerous. In the proposed research, the Chest X-Ray (CXR) images are used as an input image for experimenting the research and image processing techniques has been used to classify the nodule as benign or malignant and executed with greater accuracy in prediction and classification level. In this proposed research work, features were extracted from hasil segmentation image by using Grey Level Co- occurrence Matrix (GLCM) method. The extracted features from image are taken as input data and processed with Artificial Neural Network (ANN) Classifier. The classification and training has been done by Artificial Neural Network with back propagation (ANN-BP) method; therefore, the Artificial Neural Network has competitive and greater in executing the results by comparing with the existing methods of Linear Discriminant Analysis (LDA) and Support Vector Machine (SVM). Therefore, the performance evaluation of Artificial Neural Network has less training time with better accuracy of 87.5%, sensitivity of 97.75% and specificity of 89.75% by classifying the detected nodule as benign or malignant.