Identification of Lung Cancer Nodules from CT Images Using 2D Convolutional Neural Networks

Detection of malignant nodules at early stages from computed tomography images is time-consuming and challenging for radiologists. An alternative approach is to introduce computer-aided-diagnosis systems. Recently, deep learning approaches have outperformed other classification methods. In this paper, we use 2D convolutional neural networks to detect malignant nodules from CT scan images. We use modified VGG16 for the identification of lung cancer. LUNA 16 dataset is used to train and evaluate the proposed method, and experimental results show encouraging identification performance of the proposed method. We also compare the performance of the proposed method with the existing 2D CNN methods.