首页 > 论文 > 激光与光电子学进展 > 57卷 > 24期(pp:241025--1)

基于深度学习特征融合的视网膜图像分类

Deep Learning Feature Fusion-Based Retina Image Classification

  • 摘要
  • 论文信息
  • 参考文献
  • 被引情况
  • PDF全文
分享:

摘要

针对光学相干层析视网膜图像进行人工分类诊断时存在漏检、效率低等问题,提出一种基于深度学习技术构建联合多层特征的卷积神经网络分类算法。首先通过均值漂移和数据归一化算法对视网膜图像进行预处理,并结合损失函数加权算法解决数据不平衡问题;其次使用轻量深度可分离卷积替代普通卷积层,降低模型参数量,采用全局平均池化替换全连接层,增加空间鲁棒性,并联合不同卷积层构建特征融合层,加强层间特征流通;最后使用SoftMax分类器进行图像分类。实验结果表明,该模型在准确率、精确率、召回率上分别达到97%、95%、97%,缩短了识别时长,所提方法在视网膜图像分类诊断中具有良好的性能。

Abstract

Aiming at the problems of missed detection and low efficiency in manual classification and diagnosis of optical coherence tomography retina images, a deep learning-based convolutional network classification algorithm is proposed to construct joint multilayer features. First, retinal images are preprocessed using the mean shift and data normalization algorithm. The loss function weighting algorithm is combined to solve the data imbalance problem. Second, a lightweight deep separable convolution rather than an ordinary convolution layer is used to reduce the number of model parameters. Global average pooling replaces fully connected layers to increase spatial robustness, and different convolutional layers are used to build feature fusion layers to enhance feature circulation between layers. Finally, the SoftMax classifier is used for image classification. Experimental results show that the model can achieve 97%, 95%, and 97% in accuracy, precision, and recall, respectively, thereby reducing the recognition time. The proposed deep learning feature fusion-based method performs well in the classification and diagnosis of retinal images.

广告组1 - 空间光调制器+DMD
补充资料

中图分类号:TP391.4

DOI:10.3788/LOP57.241025

所属栏目:图像处理

基金项目:国家自然科学基金、福建省科技计划项目;

收稿日期:2020-04-24

修改稿日期:2020-06-09

网络出版日期:2020-12-01

作者单位    点击查看

张添福:福州大学机械工程及自动化学院, 福建 福州 350108
钟舜聪:福州大学机械工程及自动化学院, 福建 福州 350108
连超铭:福州大学机械工程及自动化学院, 福建 福州 350108
周宁:福州大学机械工程及自动化学院, 福建 福州 350108
谢茂松:福建医科大学附属第一医院, 福建 福州 350000

联系人作者:钟舜聪(zhongshuncong@hotmail.com)

备注:国家自然科学基金、福建省科技计划项目;

【1】Wang J R, Yang Y. Research progress of related factors of type Ⅱ diabetic retinopathy [J]. Journal of Kunming Medical University. 2019, 40(4): 131-135.
王金瑞, 杨莹. 2型糖尿病视网膜病变相关因素研究进展 [J]. 昆明医科大学学报. 2019, 40(4): 131-135.

【2】Cho N H, Shaw J E, Karuranga S, et al. IDF diabetes atlas: global estimates of diabetes prevalence for 2017 and projections for 2045 [J]. Diabetes Research and Clinical Practice. 2018, 138: 271-281.Cho N H, Shaw J E, Karuranga S, et al. IDF diabetes atlas: global estimates of diabetes prevalence for 2017 and projections for 2045 [J]. Diabetes Research and Clinical Practice. 2018, 138: 271-281.

【3】Laddha A P, Kulkarni Y A. Tannins and vascular complications of diabetes: an update [J]. Phytomedicine. 2019, 56: 229-245.

【4】Yang L, Shen X. Research on correlation between diabetic retinopathy and dry eye [J]. International Eye Science. 2018, 18(4): 744-747.
杨玲, 沈玺. 糖尿病性视网膜病变与干眼的相关性研究 [J]. 国际眼科杂志. 2018, 18(4): 744-747.

【5】Bi K, Wang Y. Advances in the application of computer aided diagnosis in ultrasound medicine [J]. Oncoradiology. 2019, 28(5): 296-300.
毕珂, 王茵. 计算机辅助诊断技术在超声医学中的应用进展 [J]. 肿瘤影像学. 2019, 28(5): 296-300.

【6】Wang C, He X X, Fang L Y, et al. Automatic classification of retinal optical coherence tomography images via convolutional neural networks with joint decision [J]. Chinese Journal of Biomedical Engineering. 2018, 37(6): 641-648.
王翀, 何兴鑫, 方乐缘, 等. 基于联合决策卷积神经网络的光学相干断层扫描图像自动分类 [J]. 中国生物医学工程学报. 2018, 37(6): 641-648.

【7】Bhowmik A, Kumar S, Bhat N. Eye disease prediction from optical coherence tomography images with transfer learning . [C]∥Communications in Computer and Information Science, May 15, 2019. Rome: ICFNN. 2019, 104-114.

【8】Yu H C. Research on classification of retinal diseases based on SE-block [D]. Changchun: Jilin University. 2019.
于海琛. 基于SE-Block的视网膜疾病分类方法研究 [D]. 长春: 吉林大学. 2019.

【9】Voets M, M?llersen K. -02-11) [2020-05-25] . https:∥arxiv. 2015, org/abs/1803: 04337.

【10】Zhang R G. Retina vessel segmentation based on multi-scale and multi-path fully convolutional neural network [J]. Laser Journal. 2020, 41(2): 194-198.
张润谷. 基于多尺度多路径FCN的视网膜血管分割 [J]. 激光杂志. 2020, 41(2): 194-198.

【11】Zhuo D, Jing J F, Zhang H H, et al. Classification of chopped strand mat defects based on convolutional neural network [J]. Laser & Optoelectronics Progress. 2019, 56(10): 101009.
卓东, 景军锋, 张缓缓, 等. 基于卷积神经网络的短切毡缺陷分类 [J]. 激光与光电子学进展. 2019, 56(10): 101009.

【12】Meng T, Liu Y H, Zhang K Y. Algorithm for pathological image diagnosis based on boosting convolutional neural network [J]. Laser & Optoelectronics Progress. 2019, 56(8): 081001.
孟婷, 刘宇航, 张凯昱. 一种基于增强卷积神经网络的病理图像诊断算法 [J]. 激光与光电子学进展. 2019, 56(8): 081001.

【13】Feng Y, Yi B S, Wu C Y, et al. Pulmonary nodule recognition based on three-dimensional convolution neural network [J]. Acta Optica Sinica. 2019, 39(6): 0615006.
冯雨, 易本顺, 吴晨玥, 等. 基于三维卷积神经网络的肺结节识别研究 [J]. 光学学报. 2019, 39(6): 0615006.

【14】Yang M J, Tang Y Q, Jiang X J. Novel shoe type recognition method based on convolutional neural network [J]. Laser & Optoelectronics Progress. 2019, 56(19): 191505.
杨孟京, 唐云祁, 姜晓佳. 基于卷积神经网络的鞋型识别方法 [J]. 激光与光电子学进展. 2019, 56(19): 191505.

【15】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.
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.

【16】Ioffe S. -03-02)[2020-05-25] . https:∥arxiv. 2015, org/abs/1502: 03167.

【17】Li K, Zou C Q, Bu S H, et al. Multi-modal feature fusion for geographic image annotation [J]. Pattern Recognition. 2018, 73: 1-14.

【18】Lecun Y, Bengio Y, Hinton G. Deep learning [J]. Nature. 2015, 521(7553): 436.

【19】Kermany D S, Goldbaum M, Cai W J, et al. Identifying medical diagnoses and treatable diseases by image-based deep learning [J]. Cell. 2018, 172(5): 1122-1131.

【20】Zhang K, Wang X, Guo Y, Image Processing, et al. Sydney, Australia . New York: IEEE:. 2019, 19297873.

引用该论文

Zhang Tianfu,Zhong Shuncong,Lian Chaoming,Zhou Ning,Xie Maosong. Deep Learning Feature Fusion-Based Retina Image Classification[J]. Laser & Optoelectronics Progress, 2020, 57(24): 241025

张添福,钟舜聪,连超铭,周宁,谢茂松. 基于深度学习特征融合的视网膜图像分类[J]. 激光与光电子学进展, 2020, 57(24): 241025

您的浏览器不支持PDF插件,请使用最新的(Chrome/Fire Fox等)浏览器.或者您还可以点击此处下载该论文PDF