A improved V-Net lung nodule segmentation method based on Res2Net
LI Yuanlong
MEN Jingru
LIU Jialin
CHEN Qi
WU Liang
Abstract：In order to improve the segmentation efficiency of nodules in lung computed tomography(CT)images,we improved the V-Net 3D segmentation model and proposed the Res2Net-Vnet.Firstly,the image was preprocessed to alleviate the influence of une-ven positive and negative samples.Then,the original single-size convolution in V-Net was replaced with Res2Net convolution,and the scale feature of the receptive field was increased while reducing the model parameters by using small convolution kernel stacks.The ex-periments on Luna16 subdataset indicated that the Dice similarity coefficients of V-Net,SKV-Net and Res2Net-Vnet was 0.703,0.750 and 0.764,respectively.The results show that Res2Net-Vnet is superior to V-Net and SKV-Net in pulmonary nodule segmenta-tion,which proves the certain effectiveness in improving pulmonary nodule segmentation.
Keywords：Lung cancerConvolutional networkMulti-scale fusionResidual connectionLightweightAuxiliary diagnosis
Publication Date：2024-10-28
Online Publishing Date：2026-09-11(First online date of this platform, not the publication date of the document)
Pages：6( 356-361 )