摘要
针对目前视网膜血管分割中存在的细小血管提取不完整、分割不准确的问题,从血管形状拓扑关系利用的角度出发,探索多任务卷积神经网络设计,提出骨架图引导的级联视网膜血管分割网络框架.该框架包含血管骨架图提取网络模块、血管分割网络模块和若干自适应特征融合结构体.骨架提取辅助任务用于提取血管中心线,能够最大限度地保留血管拓扑结构特征;自适应特征融合结构体嵌入在两个模块的特征层间.该结构体通过学习像素级的融合权重,有效地将血管拓扑结构特征与血管局部特征相融合,加强血管特征的结构信息响应.为了获得更完整的骨架图,骨架图提取网络还引入了基于图的正则化损失函数用于训练.与最新的血管分割方法相比,该方法在3个公共视网膜图像数据集上均获得第一名,在DRIVE,STARE和CHASEDB1中其F1值分别为83.1%,85.8%和82.0%.消融实验表明骨架图引导的视网膜血管分割效果更好,并且,基于图的正则化损失也能进一步提高血管分割准确性.通过将骨架提取模块和血管分割模块替换成不同的卷积网络验证了框架的普适性.
Accurate identification of retinal vessels is essential for assisting doctors in screening early fundus diseases.Diabetes,hypertension,and cardiovascular disease can cause abnormalities of the retinal vascular structure.Retinal vessel segmentation maps can be quickly obtained using the automated retinal vessel segmentation technology,which saves time and cost of manually identifying retinal vessels.Aiming at the problem of incomplete and inaccurate extraction of fine retinal vessels,this paper explored the design of a multitask convolutional neural network and the topological relationship of retinal vessels.A cascaded retinal vessel segmentation network framework guided by a skeleton map was proposed.The auxiliary task of skeleton extraction was used to extract vessel centerlines,which could maximally preserve topological structure information.SAFF cascaded the two modules by remaining embedded between their feature layers.This process could effectively fuse the structural features with the vessel local features by learning pixel-wise fusion weight and thus enhancing the structural response of features in the vessel segmentation module.To obtain a complete skeleton map,the skeleton map extraction module introduced a graph-based regularization loss function for training.Compared with the latest vessel segmentation methods,the proposed approach wins the first place among the three public retinal image datasets.F1 metrics of the proposed method achieved 83.1%,85.8%,and 82.0%on the DRIVE,STARE,and CHASEDB1 datasets,respectively.Ablation studies have shown that skeleton map-guided vessel segmentation is more effective,and graph-based regularization loss further improves accuracy of the retinal vessel segmentation compared to the vanilla network.Moreover,the framework generality is verified by replacing the skeleton map extraction and vessel segmentation modules with various convolutional networks.
作者
姜大光
李明鸣
陈羽中
丁文达
彭晓婷
李瑞瑞
JIANG Da-guang;LI Ming-ming;CHEN Yu-zhong;DING Wen-da;PENG Xiao-ting;LI Rui-rui(School of Information Science and Technology,Beijing University of Chemical and Technology,Beijing 100029,China;Beijing Airdoc Technology Co.,Ltd,Beijing 100089,China;Beijing Futong Dongfang Technology Co.,Ltd,Beijing 100010,China)
出处
《工程科学学报》
EI
CSCD
北大核心
2021年第9期1244-1252,共9页
Chinese Journal of Engineering
基金
北京化工大学‒中日友好医院生物医学转化工程研究中心联合资助项目(XK2020-7)
科技部重点研发资助项目(2020YFF0305100)。
关键词
骨架提取
视网膜血管分割
多任务
级联网络
基于图的正则化
skeleton extraction
retinal vessel segmentation
multitask
cascaded network
graph-based regularization
