Background: Accurately detecting and examining lung nodules early is key in diagnosing lung cancers and thus\none of the best ways to prevent lung cancer deaths. Radiologists spend countless hours detecting small\nspherical-shaped nodules in computed tomography (CT) images. In addition, even after detecting nodule candidates,\na considerable amount of effort and time is required for them to determine whether they are real nodules. The aim of\nthis paper is to introduce a high performance nodule classification method that uses three dimensional deep\nconvolutional neural networks (DCNNs) and an ensemble method to distinguish nodules between non-nodules.\nMethods: In this paper, we use a three dimensional deep convolutional neural network (3D DCNN) with shortcut\nconnections and a 3D DCNN with dense connections for lung nodule classification. The shortcut connections and\ndense connections successfully alleviate the gradient vanishing problem by allowing the gradient to pass quickly and\ndirectly. Connections help deep structured networks to obtain general as well as distinctive features of lung nodules.\nMoreover, we increased the dimension of DCNNs from two to three to capture 3D features. Compared with shallow\n3D CNNs used in previous studies, deep 3D CNNs more effectively capture the features of spherical-shaped nodules.\nIn addition, we use an alternative ensemble method called the checkpoint ensemble method to boost performance.\nResults: The performance of our nodule classification method is compared with that of the state-of-the-art methods\nwhich were used in the LUng Nodule Analysis 2016 Challenge. Our method achieves higher competition\nperformance metric (CPM) scores than the state-of-the-art methods using deep learning. In the experimental setup\nESB-ALL, the 3D DCNN with shortcut connections and the 3D DCNN with dense connections using the checkpoint\nensemble method achieved the highest CPM score of 0.910.\nConclusion: The result demonstrates that our method of using a 3D DCNN with shortcut connections, a 3D DCNN\nwith dense connections, and the checkpoint ensemble method is effective for capturing 3D features of nodules and\ndistinguishing nodules between non-nodules.
Loading....