基于引导滤波和Transformer的双分支视网膜血管分割网络

doi:10.6040/j.issn.1671-9352.0.2024.062

《山东大学学报(理学版)》 ›› 2025, Vol. 60 ›› Issue (7): 22-31.doi: 10.6040/j.issn.1671-9352.0.2024.062

• • 上一篇

基于引导滤波和Transformer的双分支视网膜血管分割网络

闫本聪¹,王迎美^1,2*

1. 山东理工大学数学与统计学院, 山东淄博 255049;2.山东大学数学学院, 山东济南 250100

发布日期:2025-07-01
通讯作者: 王迎美(1987— ),女,副教授,博士,研究方向为医学图像处理与重建. E-mail:yingmeiwang@sdut.edu.cn
作者简介:闫本聪(2000— ),男,硕士研究生,研究方向为医学图像处理. E-mail:1720186576@qq.com*通信作者:王迎美(1987— ),女,副教授,博士,研究方向为医学图像处理与重建. E-mail:yingmeiwang@sdut.edu.cn
基金资助:
山东省自然科学基金资助项目(ZR2022MA027);中山大学广东省计算科学重点实验室开放基金项目(2021003);山东省属普通本科高校教师访学研修项目

A dual branch retinal vessel segmentation network based on guided filtering and Transformer

YAN Bencong¹, WANG Yingmei^1,2*

1. School of Mathematics and Statistics, Shandong University of Technology, Zibo 255049, Shandong, China;
2. School of Mathematics, Shandong University, Jinan 250100, Shandong, China

Published:2025-07-01

摘要/Abstract

摘要： 引入Transformer结构和UNet++网络,提出一个新的双分支视网膜血管分割网络,该网络中的双分支编码器可以更好地关联图像中的全局信息,使得整个网络在小数据集的训练中也有较好的效果。在此基础上,为了进一步解决UNet网络中下采样操作导致的视网膜血管信息丢失问题,在输出层和网络第二层的特征图引入引导滤波,可以有效提高小血管分割精度。该网络使用DRIVE数据集(digital retinal images for vessel extraction)和CHASEDB1数据集(combined healthy and diabetic retinopathy database 1)进行实验,在精确度、灵敏度等参数上有较大的提升,并且血管分割图中正确分割出更多细小血管,总体表现出较好的效果。

关键词: UNet++, Transformer, 引导滤波, 视网膜血管分割

Abstract: Based on the Transformer structure and the UNet++ network,a new dual branch retinal vessel segmentation network is proposed. The dual branch encoder in the network can better correlate global information in the image, so that the proposed network performs well on small datasets. Furthermore, in order to solve the problem of retinal vessel information loss caused by downsampling operations in UNet, guided filtering is introduced in the feature maps of the output layer and the second layer, which effectively improves the accuracy of small vessel segmentation. The effectiveness of the network is validated by experiments on the DRIVE(digital retinal images for vessel extraction)dataset and the CHASEDB1(combined healthy and diabetic retinopathy database 1)dataset, which shows significant improvements in accuracy, sensitivity, and other parameters. In addition, more small blood vessels are visually segmented with better accuracy. All results demonstrate that the proposed method has better performance.

Key words: UNet++, Transformer, guided filtering, retinal vascular segmentation

中图分类号:

闫本聪,王迎美. 基于引导滤波和Transformer的双分支视网膜血管分割网络[J]. 《山东大学学报(理学版)》, 2025, 60(7): 22-31.

YAN Bencong, WANG Yingmei. A dual branch retinal vessel segmentation network based on guided filtering and Transformer[J]. JOURNAL OF SHANDONG UNIVERSITY(NATURAL SCIENCE), 2025, 60(7): 22-31.

参考文献

[1] ZHOU Jing, CHEN Bo. Retinal cell damage in diabetic retinopathy[J]. Cells, 2023, 12(9):1342.
[2] JENA P K, KHUNTIA B, PALAI C, et al. A novel approach for diabetic retinopathy screening using asymmetric deep learning features[J]. Big Data and Cognitive Computing, 2023, 7(1):25.
[3] YOU Ning, HAN Libo, ZHU Daming, et al. Research on image denoising in edge detection based on wavelet transform[J]. Applied Sciences, 2023, 13(3):1837.
[4] XIAN Ronghao, LUGU Rikong, PENG Hong, et al. Edge detection method based on nonlinear spiking neural systems[J]. International Journal of Neural Systems, 2023, 33(1):2250060.
[5] 李兰兰,张孝辉,牛得草,等. 深度学习在视网膜血管分割上的研究进展[J]. 计算机科学与探索,2021,15(11):2063-2076. LI Lanlan, ZHANG Xiaohui, NIU Decao, et al. Research progress of deep learning in retinal vessel segmentation[J]. Journal of Fronties of Computer Science and Technology, 2021, 15(11):2063-2076.
[6] 纪守领,杜天宇,邓水光,等. 深度学习模型鲁棒性研究综述[J]. 计算机学报,2022,45(1):190-206. JI Shouling, DU Tianyu, DENG Shuiguang, et al. Robustness certification research on deep learning models: a survey[J]. Chinese Journal of Computers, 2022, 45(1):190-206.
[7] RONNEBERGER O, FISCHER P, BROX T. U-Net: convolutional networks for biomedical image segmentation[C] // Medical Image Computing and Computer-Assisted Intervention. Cham: Springer, 2015:234-241.
[8] SIDDIQUE N, PAHEDING S, ELKIN C P, et al. U-net and its variants for medical image segmentation: a review of theory and applications[J]. IEEE Access, 2021, 9:82031-82057.
[9] ZHOU Z W, SIDDIQUEE M M R, TAJBAKHSH N, et al. Unet++: redesigning skip connections to exploit multiscale features in image segmentation[J]. IEEE Transactions on Medical Imaging, 2019, 39(6):1856-1867.
[10] LONG J, SHELHAMER E, Darrell T. Fully convolutional networks for semantic segmentation[C] // Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition. Boston: lEEE, 2015:3431-3440.
[11] VASWANI A, SHAZEER N, PARMAR N, et al. Attention is all you need[C] // Proceedings of the 31st International Conference on Neural Information Processing Systems. New York: ACM, 2017:6000-6010.
[12] DOSOVISKIY A, BEYER L, KOLESNIKOV A, et al. An image is worth 16x16 words: transformers for image recognition at scale[EB/OL].(2021-06-03)[2024-03-01]. https://arxiv.org/pdf/2010.11929.
[13] CHEN Jieneng, LU Yongyi, YU Qihang, et al. Transunet: transformers make strong encoders for medical image segmentation[EB/OL].(2021-02-08)[2024-03-01]. https://arxiv.org/pdf/2102.04306.
[14] TOMAR N K, SHERGILL A, RIEDERS B, et al. TransResU-Net: transformer based ResU-Net for real-time colonoscopy polyp segmentation[EB/OL].(2022-06-17)[2024-03-01]. https://doi.org/10.48550/arXiv.2206.08985.
[15] CAO H, WANG Y Y, CHEN J, et al. Swin-unet: unet-like pure transformer for medical image segmentation[C] // European Conference on Computer Vision. Cham: Springer, 2022:205-218.
[16] HE Kaiming, SUN Jian, TANG Xiaoou. Guided image filtering[J]. IEEE Transactions on Pattern Analysis and Machine Intelligence, 2012, 35(6):1397-1409.
[17] QIN Zhen, SUN Weixuan, DENG Hui, et al. Cosformer: rethinking softmax in attention[EB/OL].(2022-02-17)[2024-03-01]. https://doi.org/10.48550/arXiv.2206.08985.
[18] HE Kaiming, ZHANG Xiangyu, REN Shaoqing, et al. Deep residual learning for image recognition[C] // Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition. Las Vegas: IEEE, 2016:770-778.
[19] WU Yuxin, HE Kaiming. Group normalization[C] // Proceedings of the European Conference on Computer Vision(ECCV). Munich: Springer, 2018:3-19.
[20] 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.
[21] OWEN C G, RUDNICKA A R, MULLEN R, et al. Measuring retinal vessel tortuosity in 10-year-old children: validation of the computer-assisted image analysis of the retina(CAIAR)program[J]. Investigative Ophthalmology & Visual Science, 2009, 50(5):2004-2010.
[22] REYAD M, SARHAN A M, ARAFA M. A modifiedAdam algorithm for deep neural network optimization[J]. Neural Computing and Applications, 2023, 35(23):17095-17112.
[23] QURESHI I, YAN J, ABBAS Q, et al. Medical image segmentation using deep semantic-based methods:a review of techniques, applications and emerging trends[J]. Information Fusion, 2023, 90:316-352.
[24] YIN Pengshuai, YUAN Rui, CHENG Yiming, et al. Deep guidance network for biomedical image segmentation[J]. IEEE Access, 2020, 8:116106-116116.
[25] SUN Guanqun, PAN Yizhi, KONG Weikun, et al. DA-TransUNet: integrating spatial and channel dual attention with transformer U-Net for medical image segmentation[J]. Frontiers in Bioengineering and Biotechnology, 2024, 12:1398237.

多维度评价

Viewed

Full text

Abstract

Cited

Shared

Discussed

基于引导滤波和Transformer的双分支视网膜血管分割网络

A dual branch retinal vessel segmentation network based on guided filtering and Transformer

PDF (PC)

摘要/Abstract

引用本文

使用本文

参考文献

相关文章 2

多维度评价

本文评价

推荐阅读 0

[1]	易三莉,陈建亭,贺建峰. ASR-UNet: 一种基于注意力机制改进的视网膜血管[J]. 《山东大学学报(理学版)》, 2021, 56(9): 13-20.
[2]	林明星. 基于变分结构引导滤波的低照度图像增强算法[J]. 《山东大学学报(理学版)》, 2020, 55(9): 72-80.