Research on pulmonary nodule segmentation algorithm based on improved V-Net

To solve the problem that the segmentation of lung nodules in CT images is not accurate enough, a lung nodule segmentation algorithm based on an improved V-Net network is proposed. First, the network structure is improved because the original V-Net network cannot make full use of the feature map information, so that the model can make full use of CT image information. Then the combined loss function is used to prevent missed detection in the model training, which improves the convergence speed of the model. By using the LUNA16 dataset to carry out this lung nodule segmentation experiment, the Dice similarity coefficient, accuracy rate and recall rate were obtained by 0.6910, 0.8158 and 0.6525, respectively, and the experimental results showed that the algorithm can divide the lung nodules very well.