An Approach of AlexNet CNN Algorithm Model for Lung Cancer Detection and Classification

Abstract

As a reliable tool for identifying and classifying different illnesses, including lung cancer, deep learning has grown significantly in popularity. It is crucial to quickly and accurately diagnose lung cancer because different treatment options depend on the type and stage of the disease. Deep learning algorithms (DLA) are used to speed up the critical process of lung cancer detection and lessen the burden on medical professionals. In this study, the feasibility of employing deep learning algorithms for the early detection of lung cancer is explored, using data from the Lung Imaging Database Consortium (LIDC) database. The study introduces the VGG-16 and AlexNet models specifically to identify the presence of cancer in lung images. The AlexNet model is chosen for additional classification tasks based on performance. The suggested technique displays considerable increases in both the prediction and classification accuracy of cancer. The results from using the AlexNet model show the highest levels of accuracy, with classification accuracy of 97.76% and prediction accuracy of 97.02%, both verified using a 5-fold cross-validation method. Moreover, when classifying the forms of cancer, the model gets a remarkable area under the curve (AUC) value of 1 for the Adenocarcinoma class, signaling extraordinary performance. Notably, the proposed model achieves an accuracy exceeding 90% across all classes.