Abstract

Lung cancer is the major reason for numerous numbers of deaths happening all over the world. Lung cancer is difficult to identify since the symptoms only towards the end of the process. Early discovery and accurate treatment of the situation, can lower the death rate. Cancerous cells develop in the human body due to lung nodules developing in the lungs. There are two sorts of lung nodules: malignant and non-cancerous. The type of lung nodule should be detected to classify if a human has cancer or not. For that, image classification should be performed on CT scan images. A group of CT scan images are used as a dataset to build a model and classify if the patient’s CT scan has cancerous nodules or non-cancerous nodules. There exist many models which detect lung cancer with CT scan images but early detection of it is still not entirely possible. The main reason for numerous deaths is the failure of early detection. Since the symptoms do not show till the cancer becomes more severe. The aim is to identify the existence of cancerous lung nodules using patient CT images with and without lung cancer. To create an appropriate classifier, this project follows different approaches from computer vision and deep learning, specifically Convolutional Neural Networks(CNN).