计算机科学与探索 ›› 2022, Vol. 16 ›› Issue (8): 1877-1884.DOI: 10.3778/j.issn.1673-9418.2012011

• 图形图像 • 上一篇    下一篇

改进U型网络在视网膜病变检测中的应用研究

杨知桥1, 张莹1,+(), 王新杰1, 张东波1,2, 王玉1   

  1. 1. 湘潭大学 自动化与电子信息学院,湖南 湘潭 411105
    2. 机器人视觉感知与控制技术国家工程实验室,长沙 410082
  • 收稿日期:2020-12-03 修回日期:2021-01-28 出版日期:2022-08-01 发布日期:2021-02-05
  • 通讯作者: +E-mail: zhangying@xtu.edu.cn
  • 作者简介:杨知桥(1996—),男,湖南桃源人,硕士研究生,主要研究方向为医学图像处理。
    张莹(1972—),男,湖南长沙人,博士,副教授,主要研究方向为机器人控制、模式识别。
    王新杰(1995—),男,湖南郴州人,硕士研究生,主要研究方向为模式识别、图像处理、视觉导航。
    张东波(1973—),男,湖南隆回人,博士,教授,CCF会员,主要研究方向为模式识别、图像处理、深度学习。
    王玉(1996—),女,新疆阿图什人,硕士研究生,主要研究方向为模式识别、图像处理。
  • 基金资助:
    国家自然科学基金(61175075);国家自然科学基金区域创新发展联合基金(U19A2083);湖南省战略性新兴产业科技攻关与重大成果转化项目(2019GK4007);湖南省重点学科资助项目。

Application Research of Improved U-shaped Network in Detection of Retinopathy

YANG Zhiqiao1, ZHANG Ying1,+(), WANG Xinjie1, ZHANG Dongbo1,2, WANG Yu1   

  1. 1. College of Automation and Electronic Information, Xiangtan University, Xiangtan, Hunan 411105, China
    2. National Engineering Laboratory for Robotic Visual Perception and Control Technology, Changsha 410082, China
  • Received:2020-12-03 Revised:2021-01-28 Online:2022-08-01 Published:2021-02-05
  • About author:YANG Zhiqiao, born in 1996, M.S. candidate. His research interest is medical image processing.
    ZHANG Ying, born in 1972, Ph.D., associate professor. His research interests include robot control and pattern recognition.
    WANG Xinjie, born in 1995, M.S. candidate. His research interests include pattern recognition, image processing and visual navigation.
    ZHANG Dongbo, born in 1973, Ph.D., professor, member of CCF. His research interests include pattern recognition, image processing and deep learning.
    WANG Yu, born in 1996, M.S. candidate. Her research interests include pattern recognition and image processing.
  • Supported by:
    the National Natural Science Foundation of China(61175075);the Joint Fund for Regional Innovation and Development of National Natural Science Foundation of China(U19A2083);the Scientific and Technological Breakthrough and Major Achievement Transformation Project of Strategic Emerging Industries in Hunan Province(2019GK4007);and the Funding for Key Disciplines of Hunan Province.

摘要:

眼底视网膜血管分析和渗出物、出血点等主要病灶区检测是判断糖尿病性视网膜病变程度的重要方法。针对细微血管的分叉以及端点处分割效果不好、渗出物边界不明显以及出血点细小且分布零散不易分割等问题,提出一种改进U型网络,通过改进上下文提取编码模块,提取更丰富的高级别特征;并在特征编码阶段加入混合注意力机制(HAM),突出细微血管以及病灶区特征,减小背景类和噪声影响。实验结果表明,提出的算法在眼底视网膜血管分割数据集DRIVE上的分割准确率、灵敏度、特异性和AUC值比U-NET、CE-NET等现有方法有一定提升,其中灵敏度相较CE-Net网络提升了0.014 6。在糖尿病性视网膜病变病灶区分割数据集DIARETDB1上,对渗出物和出血点的分割效果比U-NET、CE-NET等现有方法有较好的提升,能有效辅助医生诊断。

关键词: 眼底视网膜血管, 渗出物, 出血点, 上下文特征编码, 注意力机制

Abstract:

Fundus retinal blood vessel analysis and detection of exudates and bleeding points are important methods for judging the degree of diabetic retinopathy. Aiming at the problems such as poor segmentation effect of bifurcation and end points of microvessels, unclear exudate boundary, difficult segmentation of small and scattered bleeding points, an improved U-shaped network is proposed to extract more rich high-level features by improving the context extraction coding module. And in the feature encoding stage, a hybrid attention mechanism (HAM) is added to highlight the features of microvessels and lesions, and reduce the impact of background and noise. Experimental results show that the segmentation accuracy, sensitivity, specificity and AUC value of the proposed algorithm on the fundus retinal blood vessel segmentation dataset DRIVE are better than U-NET, CE-NET and other existing methods. The sensitivity is increased by 0.0146 compared with CE-Net network. On diabetic retinopathy lesion segmentation dataset DIARETDB1, the segmentation effect of exudates and bleeding points is better than U-NET, CE-NET and other existing methods, which can effectively assist doctors in diagnosis.

Key words: fundus retinal blood vessel, exudate, bleeding point, context feature encoder, attention mechanism

中图分类号: