Lung cancer detection and classification using machine learning

Abstract

Lung cancer is a term known to all nowadays. This disease grows in the lung tissues and starts to spread with time. The cells responsible for air passage are corrupted by it. It can happen because of air pollution. When we breathe in polluted air regularly, our lungs are likely to be damaged. But by smoking, a lot of people are damaging their lungs repeatedly. Due to this act, they are receiving lung cancer as consequence. It has been affecting people acutely and if prevented in earlier states, then the rate of death would lessen. In order to do that, we have proposed some methods to detect this illness. Machine Learning is a technique where machines (computers) can give us a solution to a problem by analyzing the collected data. Using this method, we can detect lung cancer which is the first step towards our desired goal. Usage of CT scan could help us decide between cancer affected and unaffected human cells. Those cells also can be classified more efficiently and we can accurately detect the stage of the cancer when we use CNN models like VGG-19, ResNet50, EfficientNet, DenseNet and so on. We got the highest accuracy from ResNet50 which is 89.52%.