Lung Cancer Detection Using Convolutional Neural Networks from Computed Tomography Images

Abstract

Lung Cancer is one of the most prevalent diseases that is affected in its early stages to improve the chance that patients will survive is lung disease. In this study, a Convolutional Neural Network (CNN) algorithm is proposed to detect abnormal lung tissue growth. The CNN algorithm is based on the textural characteristics of the images, which enables the classification of normal and malignant images. To improve the accuracy of cancer cell detection, an automated tool using VGG-16 techniques is required. The proposed algorithm is trained using lung images from both healthy and malignant individuals and databases have been developed for various views of the CT scanning system, such as axial, coronal, and sagittal. The region proposal network and the classifier network use the VGG-16 architecture as their base layer. The algorithm's classification and detection accuracy is 98%. A quantitative analysis of the proposed network is conducted based on confusion matrix computation and classification accuracy results. The proposed CNN algorithm can assist in the early detection of lung cancer and improve the accuracy of medical diagnosis, ultimately saving precious human lives.