LUNG CANCER DETECTION USING CNN

Abstract

Abstract—Lung cancer is the major cause of death. Therefore there is a need for the importance of detection of cancer in the earlier stages to reduce the mortality rate. Previously detection is based on computed tomography (CT) scans and this cost radiologists much time to detect the tumor region. So, a DLbased model is required in the detection of lung cancer on CT scan images. This method has the advantages of accessibility, cost-effectiveness, and low radiation dose. In this project, we proposed a fully automatic method (using DL) for lung cancer detection. In the first stage, lung regions are extracted from the CT image, and in that region, each slice is segmented to get tumors. The tumor regions are first segmented and then used to train CNN architecture (here we used resnet50 and vgg16).CNN takes the CT scan image’s pixel data, trains the model, then extracts the features automatically for better classification. Index Terms—Lung Cancer Detection, CT Scan,2D CNN,Resnet50, Vgg16