引入解耦残差自注意力的边界交叉监督语义分割网络  

Boundary-cross supervised semantic segmentation network with decoupled residual self-attention

作　　者：姜坤元 李小霞[1,2] 王利 曹耀丹 张晓强 丁楠 周颖玥 JIANG Kunyuan;LI Xiaoxia;WANG Li;CAO Yaodan;ZHANG Xiaoqiang;DING Nan;ZHOU Yingyue(School of Information Engineering,Southwest University of Science and Technology,Mianyang Sichuan 621010,China;Sichuan Industrial Autonomous and Controllable Artificial Intelligence Engineering Technology Research Center,Mianyang Sichuan 621010,China;Mianyang 404 Hospital,Mianyang Sichuan 621000,China)

机构地区：[1]西南科技大学信息工程学院,四川绵阳621010 [2]四川省工业自主可控人工智能工程技术研究中心,四川绵阳621010 [3]绵阳四〇四医院,四川绵阳621000

出　　处：《计算机应用》2025年第4期1120-1129,共10页journal of Computer Applications

基　　金：国家自然科学基金资助项目(62071399,62201479);四川省科技计划项目(2023YFG0262)。

摘　　要：针对内镜语义分割网络中病灶边缘信息丢失和大面积病灶分割不全的问题,提出一种引入解耦残差自注意力(DRA)的边界交叉监督语义分割网络(BCS-SegNet)。首先,引入DRA,以增强网络对远距离关联性病灶的学习能力;其次,构建跨级交叉融合(CLF)模块,从而将编码结构中的多级特征图逐对组合,进而实现在低计算成本下图像细节与语义信息的融合;最后,使用多方向多尺度的二维Gabor变换提取边缘信息,并使用空间注意力加权特征图中的边缘特征,以监督分割网络的解码过程,从而在像素级别上提供更精准的类内分割一致性。实验结果表明,在ISIC2018皮肤镜和Kvasir-SEG/CVC-ClinicDB结肠镜数据集上,BCS-SegNet的平均交并比(mIoU)和Dice系数分别为84.27%、90.68%和79.24%、87.91%;在自建食管内镜数据集上,BCS-SegNet的mIoU和Dice系数分别为82.73%和90.84%,mIoU相较于U-net和UCTransNet分别提升了3.30%和4.97%。可见,所提网络可以达到更完整的分割区域和更清晰的边缘细节等视觉效果。Focused on the challenges of edge information loss and incomplete segmentation of large lesions in endoscopic semantic segmentation networks,a Boundary-Cross Supervised semantic Segmentation Network(BCS-SegNet)with Decoupled Residual Self-Attention(DRA)was proposed.Firstly,DRA was introduced to enhance the network’s ability to learn distantly related lesions.Secondly,a Cross Level Fusion(CLF)module was constructed to combine multi-level feature maps within the encoding structure in a pairwise way,so as to realize the fusion of image details and semantic information at low computational cost.Finally,multi-directional and multi-scale 2D Gabor transform was utilized to extract edge information,and spatial attention was used to weight edge features in the feature maps,so as to supervise decoding process of the segmentation network,thereby providing more accurate intra-class segmentation consistency at pixel level.Experimental results demonstrate that on ISIC2018 dermoscopy and Kvasir-SEG/CVC-ClinicDB colonoscopy datasets,BCSSegNet achieves the mIoU(mean Intersection over Union)and Dice coefficient of 84.27%,90.68%and 79.24%,87.91%,respectively;on the self-built esophageal endoscopy dataset,BCS-SegNet achieves the mIoU of 82.73%and Dice coefficient of 90.84%,while the above mIoU is increased by 3.30%over that of U-net and 4.97%over that of UCTransNet.It can be seen that the proposed network can realize visual effects such as more complete segmentation regions and clearer edge details.

关 键 词：食管内镜图像 医学图像分割 自注意力机制 二维Gabor变换 多尺度边缘特征 

分 类 号：TP391.4[自动化与计算机技术—计算机应用技术]