北京生物医学工程

一种眼底图像出血点的检测算法

Algorithm of hemorrhages in fundus images

作者：周梦颖杨晓宇邱媛杨春兰刘冰

单位：首都医科大学附属北京同仁医院（北京 100730） 北京工业大学环境与生命学部（北京 100124） 北京工业大学校医院眼科（北京 100124） 通信作者：周梦颖。E-mail：marine1103@126.com

关键词：糖尿病视网膜病变；眼底图像；出血点检测；灰度检测；形态学重构

分类号：R318.04

出版年·卷·期（页码）：2022·41·3（255-259）

摘要：

目的提出一种基于灰度检测和形态学重构的出血点（hemorrhages , HA）自动检测算法，以提高糖尿病视网膜病变(diabetic retinopathy，DR)眼底图像的质量和灵敏度。方法对预处理后的图像进行灰度阈值分割，保留并提取出HA和血管特征，再利用形态学方法去除血管并消除图像边缘假阳性区域，形成新算法。用新算法测试公开数据库DIARETED1中的50幅图像（45幅HA病变图像，5幅正常图像），与专家人工判断结果进行比对验证。结果该算法的灵敏度（sensitivity,SE）和特异性(specificity,SP)分别为93.33%和80.00%。结论该算法可提升眼底图像质量和灵敏度，在不借助医生经验的条件下完成快速判定，很大程度提高了筛查的效率。

Objective To propose a hemorrhages (HA) automatic detection algorithm based on gray level detection and morphological reconstruction ，and to improve the quality and sensitivity of diabetes diabetic retinopathy (DR) fundus images. Methods The preprocessed image was segmented by gray threshold, the HA and vascular features were retained and extracted, and then the morphological method was used to remove the blood vessels and eliminate the false-positive area at the edge of the image to form a new algorithm. The new algorithm was used to test 50 images (45 HA lesion images and 5 normal images) in the public database DIARETED1, and compared with the expert manual judgment results for verification. Results The sensitivity and specificity of this algorithm were 93.33% and 80.00%, respectively.Conclusions This algorithm can improve the quality and sensitivity of fundus images, and can be used to complete rapid judgment without the help of doctors' experience, which greatly improve the efficiency of screening.

参考文献：

[1]陈萌梦. 糖尿病视网膜图像自动筛查和分析的研究[D].重庆：重庆医科大学, 2014.

Chen MM. Automatic screening and analysis study on diabetic retinopathyimages[D].Chongqing：Chongqing Medical University, 2014.

[2]Soares JVB, Leandro JJG, Cesar JrRM, et al. Retinal vessel segmentation using the2-DGabor wavelet and supervised classification[J]. IEEE Transactions on Medical?? Imaging, 2006,25(9): 1214-1222.

[3]Fleming AD, Goatman KA,Philip S, et al. Automaticdetection of retinalanatomytoassist diabetic retinopathy screening [J]. Physics in Medicine &Biology. 2007, 52(2):331-345.

[4]DevarajD, Nagaveena. Detection of red lesion in diabetic retinopathy using adaptivethresholding method [J]. International Journal of Engineering Research & Technology, 2013,2(4):1889-1892．

[5]Bharali P, Medhi JP, Nirmala SR. Detection of hemorrhages in diabetic retinopathy??????? analysisusing color fundus images[C]. IEEE International Conference on Recent Trends in Information Systems. India：IEEE，2015: 237-242.

[6]Gardner GG, Keating D, Williamson T H, et al. Automatic detection of diabetic retinopathy using an artificial neural network: A screening tool[J]. British Journal of Ophthalmology, 1996, 80(11): 940-944.

[7]Karkuzhali S, Manimegalai D. Retinal hemorrhages segmentation using improvedtoboggan segmentation algorithm in diabetic retinopathy images[J]. Biomedical Research, 2018, (1): 1.

[8]Sreng S, Maneerat N, Isarakorn D, et al. Automatic hemorrhages detectionbased on fundus?????? images[C].InternationalConferenceon InformationTechnology and Electrical Engineering,Thailand: IEEE, 2015:253-257.

[9]Nutnaree K, Smith G, Bunyarit U. Automated retinal hemorrhage detection usingmorphological top hat and rule-based classification [C]. 3rd International Conference onIntelligent Computational Systems. Singapore: IEEE, 2013: 39-43

[10]Adarsh P, Jeyakumari D. Multiclass SVM-Based Automated DiagnosisofDiabeticRetinopathy[C].International Conference on Communications & Signal Processing.India: IEEE , 2013:206-210.

[11]卢旭.视网膜眼底图像中出血和血管检测算法研究[D].上海：上海交通大学, 2019.

Lu X.Research on hemorrhage and blood vessel detection in fundus images[D]. Shanghai：ShanghaiJiao Tong University, 2019

[12] Gulshan V, Peng L, Coram M, et al. Development and validation of a deep learning algorithm for detection of diabetic retinopathy in retinal fundus photographs[J]. The Journal of the American Medical Association, 2016, 316(22):2402-2410.

[13] Van Grinsven M, Van Ginneken B, Hoyng C, et al. Fast convolutional neural network training???????????????????????????????????????????????????????????????????????????????????????????????????????????????????????? using selective data sampling: Application to hemorrhage detection in color fundus images[J]. IEEE Transactions on Medical Imaging, 2016, 35(5): 1273-1284.

[14]肖志涛，赵北方，张芳，等．基于k均值聚类和自适应模板匹配的眼底出血点检测方法[J]．中国生物医学工程学报，2015，34( 3) : 264-271．

Xiao ZT, Zhao BF, Zhang F, et al． Hemorrhages detection in fundus image based on k-means clustering and adaptive template matching[J]. Chinese Journal of Biomedical Engineering, 2015,34( 3) : 264-271．

[15] Zhou L, Li P, Yu Q，et al. Automatic hemorrhage detection in color fundus images based on gradual removal of vascular branches[C]. InternationalConference on Image Processing. USA: IEEE, 2016.

[16] Sathananthavathi V，Indumathi G，Rajalakshmi R.Abnormalities detection inretinal fundus images[C]. International Conference on Inventive Communication &Computational Technologies. India:IEEE, 2017.

[17]Kauppi T, Kalesnykiene V, Kamarainen JK, et al. DIARETDB1 diabetic retinopathy database and evaluation protocol[J]. Medical Image Understanding and Analysis, 2007, 61: 1-10.

[18] Javitt JC, Canner JK, Frank RG, et al. Detecting and treating retinopathy in patients with? TypeⅠdiabetes mellitus: A healthy policy model[J]. Ophthalmology，1990, 97(4): 483-495.

服务与反馈：

【文章下载】【加入收藏】

提示：您还未登录，请登录！点此登录