| 318 | 2 | 875 |
| 下载次数 | 被引频次 | 阅读次数 |
视网膜血管的形态学变化对早期眼科疾病的诊断具有重要意义，除眼科疾病外，糖尿病、心血管疾病等同样可以通过视网膜血管的形态判别疾病进展。然而，视网膜血管本身具有复杂的组织结构，且易受到光线等因素的影响，对其准确分割并不容易。针对上述问题，提出了一种视网膜血管分割网络。该网络中首先设计了粗糙注意力融合模块(rough attention fusion module,RAFM)，该模块基于粗糙集上下近似理论，利用全局最大池化与全局平均池化对注意力系数进行上下限描述，并串行融合通道注意力机制与空间注意力机制；然后，将粗糙注意力融合模块融入Group Transformer U network (GT U-Net)，构建一种基于粗糙注意力融合机制与Group Transformer的视网膜血管分割网络；最后，基于公开DRIVE彩色眼底图像数据集进行对比实验，该网络结构在测试集上的准确率、F1分数、AUC值分别达到了0.963 1、0.848 8和0.981 2，与GT U-Net模型相比，F1分数、AUC值分别提升了0.35%、0.21%；与其他当前主流的视网膜血管分割网络进行对比，具有一定优势。
Abstract：The morphological changes in retinal vessels play a crucial role in the diagnosis of early ophthalmic diseases. Beyond eye diseases, conditions such as diabetes and cardiovascular diseases can also be identified through the morphology of retinal vessels. However, retinal vessels possess a complex tissue structure and are easily influenced by factors such as lighting, making their accurate segmentation challenging. To address these issues, a retinal vessel segmentation network that initially incorporates a rough attention fusion module(RAFM) is proposed. This module is based on the theory of rough set upper and lower approximations, employing global max pooling and global average pooling to describe the upper and lower bounds of attention coefficients, and sequentially integrates channel attention mechanisms with spatial attention mechanisms. Subsequently, the RAFM is integrated into the Group Transformer U network(GT U-Net), constructing a retinal vessel segmentation network based on the rough attention fusion mechanism and Group Transformer. Finally, comparative experiments conducted on the publicly available DRIVE color fundus image dataset demonstrate that the network achieves an accuracy, F1 score, and AUC of 0.963 1, 0.848 8, and0.981 2, respectively, on the test set. Compared to the GT U-Net model, the F1 score and AUC were improved by 0.35%and 0.21%, respectively; and when compared to other contemporary mainstream retinal vessel segmentation networks, it exhibits certain advantages.
[1]梅旭璋，江红，孙军.基于密集注意力网络的视网膜血管图像分割[J].计算机工程，2020, 46(3):267-272.MEI X Z, JIANG H, SUN J. Retinal vessel image segmentation based on dense attention network[J]. Computer Engineering, 2020, 46(3):267-272.(in Chinese)
[2] JIN Q G, MENG Z P, PHAM T D, et al. DUNet:a deformable network for retinal vessel segmentation[J]. Knowledge-Based Systems, 2019, 178:149-162.
[3] LI X, JIANG Y C, LI M L, et al. Lightweight attention convolutional neural network for retinal vessel image segmentation[J]. IEEE Transactions on Industrial Informatics,2021, 17(3):1958-1967.
[4]蒋芸，刘文欢，梁菁.联合注意力和Transformer的视网膜血管分割网络[J].计算机工程与科学，2022, 44(11):2037-2047.JIANG Y, LIU W H, LIANG J. Retinal vessel segmentation network with joint attention and Transformer[J]. Computer Engineering&Science, 2022, 44(11):2037-2047.(in Chinese)
[5] RONNEBERGER O, FISCHER P, BROX T. U-net:convolutional networks for biomedical image segmentation[C]//Proceedings of the 18th International Conference on Medical Image Computing and Computer-Assisted Intervention,October 5-9, 2015, Munich, Germany. Cham:Springer,2015:234-241.
[6] XIAO X, LIAN S, LUO Z M, et al. Weighted res-U Net for high-quality retina vessel segmentation[C]//Proceedings of the 2018 9th International Conference on Information Technology in Medicine and Education(ITME), October 19-21, 2018, Hangzhou, China. New York:IEEE Xplore,2018:327-331.
[7]吴晨玥，易本顺，章云港，等.基于改进卷积神经网络的视网膜血管图像分割[J].光学学报，2018, 38(11):1111004.WU C Y, YI B S, ZHANG Y G, et al. Retinal vessel image segmentation based on improved convolutional neural network[J]. Acta Optica Sinica, 2018, 38(11):1111004.(in Chinese)
[8] VASWANI A, SHAZEER N, PARMAR N, et al. Attention is all you need[EB/OL].(2017-12-06)[2023-02-21].https：//arxiv. org/abs/1706. 03762v4.
[9]傅励瑶，尹梦晓，杨锋.基于Transformer的U型医学图像分割网络综述[J/OL].(2022-07-12)[2023-02-21].https：//kns.cnki.net/kcms/detail/51.1307.TP.20220711.1509.012.html.DOI:10.11772/j.issn.1001-9081.202204 0530.
[10] DOSOVITSKIY A, BEYER L, KOLESNIKOV A, et al.An image is worth 16×16 words:Transformers for image recognition at scale[EB/OL].(2021-06-03)[2023-02-21].https：//arxiv.org/abs/2010.11929.
[11] ZHANG Y D, LIU H Y, HU Q. TransFuse:fusing transformers and CNNs for medical image segmentation[C]//Proceedings of the 24th International Conference on Medical Image Computing and Computer-Assisted Intervention,September 27-October 1, 2021, Strasbourg, France. Cham:Springer, 2021:14-24.
[12] CHEN J N, LU Y Y, YU Q H, et al. TransUNet:Transformers make strong encoders for medical image segmentation[EB/OL].(2021-02-08)[2023-02-21]. https：//arxiv.org/abs/2102.04306.
[13] OKTAY O, SCHLEMPER J, LE FOLGOC L, et al. Attention U-net:learning where to look for the pancreas[EB/OL].(2018-05-20)[2023-02-21]. https：//arxiv.org/abs/1804.03999.
[14] GU R, WANG G T, SONG T, et al. CA-net:comprehensive attention convolutional neural networks for explainable medical image segmentation[J]. IEEE Transactions on Medical Imaging, 2021, 40(2):699-711.
[15] YUAN Y C, ZHANG L, WANG L T, et al. Multi-level attention network for retinal vessel segmentation[J]. IEEE Journal of Biomedical and Health Informatics, 2022, 26(1):312-323.
[16] LI Y X, WANG S, WANG J, et al. GT U-net:a U-net like group transformer network for tooth root segmentation[C]//Proceedings of the 12th International Workshop on Machine Learning in Medical Imaging, MLMI 2021, September 27, 2021, Strasbourg, France. Cham:Springer, 2021:386-395.
[17]王国胤，姚一豫，于洪.粗糙集理论与应用研究综述[J].计算机学报，2009, 32(7):1229-1246.WANG G Y, YAO Y Y, YU H. A survey on rough set theory and applications[J]. Chinese Journal of Computers,2009, 32(7):1229-1246.(in Chinese)
[18] PAWLAK Z. Rough sets[J]. International Journal of Computer&Information Sciences, 1982, 11(5):341-356.
[19]胡可云，陆玉昌，石纯一.粗糙集理论及其应用进展[J].清华大学学报（自然科学版），2001, 41(1):64-68.HU K Y, LU Y C, SHI C Y. Advances in rough set theory and its appliations[J]. Journal of Tsinghua University(Science and Technology), 2001, 41(1):64-68.(in Chinese)
[20] JIANG Z H, YU W H, ZHOU D Q, et al. ConvBERT:improving BERT with span-based dynamic convolution[EB/OL].(2020-08-06)[2023-02-21]. https://arxiv.org/abs/2008.02496.
[21] LIU Z, LIN Y T, CAO Y, et al. Swin Transformer:hierarchical Vision Transformer using Shifted Windows[C]//Proceedings of the 2021 IEEE/CVF International Conference on Computer Vision(ICCV), October 10-17, 2021,Montreal, QC, Canada. New York:IEEE Xplore, 2021:9992-10002.
[22] HU J, SHEN L, ALBANIE S, et al. Squeeze-and-excitation networks[J]. IEEE Transactions on Pattern Analysis and Machine Intelligence, 2020, 42(8):2011-2023.
[23] WOO S, PARK J, LEE J Y, et al. CBAM:convolutional block attention module[C]//Proceedings of the European Conference on Computer Vision, September 8-14, 2018,Munich, Germany. Cham:Springer, 2018:3-19.
[24] FU J, LIU J, TIAN H J, et al. Dual attention network for scene segmentation[C]//Proceedings of the 2019 IEEE/CVF Conference on Computer Vision and Pattern Recognition(CVPR), June 15-20, 2019, Long Beach, CA, USA.New York:IEEE Xplore, 2019:3141-3149.
[25] MEI H, ZHANG H, JIANG Z. Self-attention fusion module for single remote sensing image super-resolution[C]//Proceedings of the 2021 IEEE International Geoscience and Remote Sensing Symposium IGARSS, July 11-16,Brussels, Belgium. New York:IEEE Xplore, 2021:2883-2886.
[26] STAAL J, ABRàMOFF M D, NIEMEIJER M, et al.Ridge-based vessel segmentation in color images of the retina[J]. IEEE Transactions on Medical Imaging, 2004,23(4):501-509.
[27] ALOM M Z, HASAN M, YAKOPCIC C, et al. Recurrent residual convolutional neural network based on U-Net(R2UNet)for medical image segmentation[EB/OL].(2018-05-29)[2023-02-21]. https：//arxiv.org/abs/1802.06955.
[28] AZAD R, ASADI-AGHBOLAGHI M, FATHY M, et al.Bi-directional ConvLSTM U-net with densley connected convolutions[C]//Proceedings of the 2019 IEEE/CVF InternationalConference on Computer Vision Workshop(ICCVW),October 27-28, 2019, Seoul, Korea(South). New York:IEEE Xplore, 2019:406-415.
[29]孙颖，丁卫平，黄嘉爽，等. RCAR-UNet：基于粗糙通道注意力机制的视网膜血管分割网络[J].计算机研究与发展，2023, 60(4):947-961.SUN Y, DING W P, HUANG J S, et al. RCAR-U Net:retinal vessels segmentation network based on rough channel attention mechanism[J]. Journal of Computer Research and Development, 2023, 64(4):947-961.(in Chinese)
基本信息:
DOI：10.12194/j.ntu.20230306001
中图分类号:TP391.41;R774.1
引用信息:
[1]王海鹏,高自强,董佳俊,等.基于粗糙注意力融合机制与Group Transformer的视网膜血管分割网络[J].南通大学学报(自然科学版),2024,23(01):28-37，48.DOI:10.12194/j.ntu.20230306001.
基金信息:
国家自然科学基金面上项目（61976120）; 江苏省自然科学基金面上项目（BK20231337）; 江苏省高校重大自然科学基金项目（21KJA510004）; 江苏省研究生科研与实践创新计划项目（SJCX22＿1615）; 国家级大学生创新创业训练计划项目（202210304030Z）
2023-04-17
2023-04-17
2023-04-17