CADCNet:一种改进的视网膜血管分割算法
DOI:
作者:
作者单位:

1.山东理工大学数学与统计学院 淄博;2.山东新华医疗器械股份有限公司 淄博

作者简介:

通讯作者:

中图分类号:

TP.41

基金项目:

山东省自然科学基金面上项目(No. ZR2022MA027);中山大学广东省计算科学重点实验室开放基金(2021003);横向课题(2D-C-20190158)第一作者岳昱超(1998—),山东威海,硕士研究生,研究方向医学图像处理。Email: yueyuchao1998@163.com。通信作者秦嘉川(1971—),江苏无锡,高级工程师,研究方向医学影像与技术。Email: qinjiachuan@shinva.com。


CADCNet:an improved algorithm for retinal vessel segmentation
Author:
Affiliation:

School of Mathematics and Statistics,Shandong University of Technology

Fund Project:

  • 摘要
  • |
  • 图/表
  • |
  • 访问统计
  • |
  • 参考文献
  • |
  • 相似文献
  • |
  • 引证文献
  • |
  • 资源附件
  • |
  • 文章评论
    摘要:

    目前传统的视网膜血管分割方法存在的视盘混淆引起的误分割、分割结果缺乏连续性,以及细节区域分割不精准等问题,为解决这一难题,本文提出了一种基于UNet的视网膜血管分割算法。该算法利用两个水平和垂直一维卷积和二维方形卷积的融合替代传统方形卷积,提高了眼球区域的表征能力;采用了多尺度分支增加特征空间的多样性,提升了网络的特征学习和表达能力。此外,为进一步改善分割效果,本文还将多层膨胀卷积引入自编码器的深层结构,替代了传统的简单池化操作,增大卷积核的大小,扩大了感受野范围,实现了多尺度浅层特征和深层特征信息的融合。本文算法在公开DRIVE和CHASE_DB1两个数据集上进行了评估,实验结果表明,本文算法的精确率和F1值上分别达到了0.9568、0.9598和0.8326、0.8304。与传统的UNet和近期部分UNet改进网络视网膜血管分割方法相比,本文算法在准确率、敏感度、特异性、F1指标上表现出一定的优势。这一验证结果充分证明了本文所提出的模型在分割任务上具备较强的精确分割能力。

    Abstract:

    Traditional retinal vessel segmentation methods often face challenges such as missegmentation caused by optic disc confusion, lack of continuity in segmentation results, and imprecise segmentation in detailed regions. To address these issues, a retinal vessel segmentation algorithm is proposed based on UNet. The algorithm replaces traditional square convolutions with a fusion of horizontal and vertical one-dimensional convolutions and two-dimensional square convolutions, enhancing the representation capability of the eye region. A multi-scale branch approach was adopted to increase feature space diversity, thereby improving the network’s feature learning and expression capabilities. Additionally, to further enhance segmentation performance, multi-layer dilated convolutions was introduced into the deep structure of the autoencoder, replacing traditional simple pooling operations. This approach enlarged the convolution kernel size and expanded the receptive field, achieving a fusion of multi-scale shallow and deep feature information. The proposed algorithm was evaluated on the public DRIVE and CHASE_DB1 datasets. Experimental results demonstrates that the algorithm achieves Precision (0.9568 and 0.9598) and F1 scores (0.8326 and 0.8304), respectively. Compared with traditional UNet and recent UNet-based retinal vessel segmentation methods, the proposed algorithm shows advantages in Accuracy, Sensitivity, Specificity, and F1 metrics. These validation results fully demonstrate the proposed model's strong capability in precise segmentation tasks.

    参考文献
    相似文献
    引证文献
引用本文

岳昱超,王迎美,秦嘉川. CADCNet:一种改进的视网膜血管分割算法[J]. 科学技术与工程, , ():

复制
分享
文章指标
  • 点击次数:
  • 下载次数:
  • HTML阅读次数:
  • 引用次数:
历史
  • 收稿日期:2023-11-27
  • 最后修改日期:2024-06-07
  • 录用日期:2024-06-24
  • 在线发布日期:
  • 出版日期:
×
亟待确认版面费归属稿件,敬请作者关注
《科学技术与工程》入选维普《中文科技期刊数据库》自然科学类期刊月度下载排行榜TOP10