Classification of benign and malignant lung nodules based on residuals and 3D VNet network

This paper proposes a classification method of lung nodules combining 3D VNet network and residual network. This method first preprocesses the original CT image and sends it to the VNet network for pulmonary nodules detection. Then the detection results are used to complete the classification of benign and malignant lung nodules through the residual network. Finally, the classification results are output. A large number of experiments have shown that this method can well extract the position information of lung nodules in CT images. Based on the LUNA16 data set, this method has achieved high accuracy, sensitivity, and specificity. Numerically, it has achieved better classification results than traditional network models.