Binary Classification Of Lung Nodules Using Two Dimensional Neural Networks

Abstract

The overlapping symptoms of lung cancer with other pulmonary diseases is one of the critical reasons that makes it difficult to differentiate the lung nodule as cancerous or non-cancerous. Computed Aided Design (CAD) systems are rigorously researched for enhancing the accuracy of the model to predict the nature of the detected nodule. The differentiation of a lung nodule into cancer or non-cancer, and nature as benign or malignant, are important aspects of analysis. The increasing use of technologies like deep learning and neural networks has proven to be the drivers of 12st century in all the domains of research. In the comprehensive study, we present a two-dimensional convolutional neural network for predicting the nodule. The accuracy of the presented model is 99% with the split ratio of the training into the ratio of 0.7:0.1:0.1.