Retinal vessel segmentation algorithm based on hybrid phase feature

doi:10.11772/j.issn.1001-9081.2017123045

Abstract

Abstract: Focusing on the issue that the phase consistency feature is deficient in detection of vascular center, a new retinal vessel segmentation algorithm based on hybrid phase feature was proposed. Firstly, an original retinal image was preprocessed. Secondly, every pixel was represented by a 4-D vector composed of Hessian matrix, Gabor transformation, Bar-selective Combination Of Shifted FIlter REsponses (B-COSFIRE) and phase feature. Finally, Support Vector Machine (SVM) was used for pixel classification to realize the segmentation of retinal vessels. Among the four features, phase feature was a new hybrid phase feature formed by phase consistency feature and Hessian matrix feature through wavelet fusion. This new phase feature not only preserves good vascular edge information by phase consistency feature, but also compensates for the deficient detection of vascular center by phase consistency feature. The average Accuracy (Acc) of the proposed algorithm evaluated on the Digital Retinal Images for Vessel Extraction (DRIVE) database is 0.9574, and the average Area Under receiver operating characteristic Curve (AUC) is 0.9702. In the experiment of using single feature for vessel extraction through pixel classification, compared with using phase consistency feature, using hybrid phase feature for vessel extraction improves the average Accuracy (Acc) from 0.9191 to 0.9478, the AUC from 0.9359 to 0.9702. The experimental results show that hybrid phase feature is more suitable for retinal vessel segmentation based on pixel classification than phase consistency feature.

Key words: retinal, vessel segmentation, feature extraction, Support Vector Machine (SVM), wavelet image fusion

摘要： 针对相位一致性特征对血管中心检测不足问题，提出基于融合相位特征的眼底视网膜血管分割算法。首先，预处理原始的视网膜图像；然后，对图像中每个像素构造4D的特征向量（包括Hessian矩阵、Gabor变换、条带选择组合位移滤波响应（B-COSFIRE）滤波、相位特征）；最后，采用支持向量机（SVM）进行像素分类，实现眼底视网膜血管的分割。其中，相位特征是将分别提取的相位一致性特征与Hessian矩阵特征进行小波融合后得到的一种新的融合相位特征。该特征既保留了相位一致性特征良好的血管边缘信息，又克服了相位一致性特征对血管中心检测的不足。在用于血管提取的数字视网膜图像（DRIVE）数据库上测得基于融合相位特征的视网膜血管分割算法的平均准确率（Acc）为0.9574，平均受试者工作曲线面积（AUC）为0.9702；且在单一特征进行像素分类提取血管的实验中，与使用相位一致性特征相比，使用融合相位特征进行像素分类提取血管的Acc由0.9191提高到0.9478，AUC由0.9359提高到0.9578。实验结果表明，融合相位特征比相位一致性特征更适用于基于像素分类的眼底视网膜血管分割算法。

关键词: 视网膜, 血管分割, 特征提取, 支持向量机, 小波图像融合

CLC Number:

TP391.413

LI Yuanyuan, CAI Yiheng, GAO Xurong. Retinal vessel segmentation algorithm based on hybrid phase feature[J]. Journal of Computer Applications, 2018, 38(7): 2083-2088.

李媛媛, 蔡轶珩, 高旭蓉. 基于融合相位特征的视网膜血管分割算法[J]. 计算机应用, 2018, 38(7): 2083-2088.

References

[1] 王昕,张歆雅,李颖昉.视网膜血管分割算法研究[J].吉林大学学报(信息科学版),2016,34(5):588-593.(WANG X, ZHANG X Y, LI Y F. Research on segmentation method of retinal vessels[J]. Journal of Jilin University (Information Science Edition), 2016, 34(5):588-593).
[2] 殷本俊,陈燕,李华婷,等.基于Morlet小波变换的视网膜血管分割[J].计算机辅助设计与图形学学报,2015,27(7):1263-1270.(YIN B J, CHEN Y, LI H T. et al. Using Morlet wavelet for retinal vessel segmentation[J]. Journal of Computer-Aided Design & Computer Graphics, 2015, 27(7):1263-1270.)
[3] 梁礼明,黄朝林,朱莎,等.基于Hessian矩阵和水平集的视网膜血管分割[J].科学技术与工程,2016,16(10):44-49.(LIANG L M, HUANG C L, ZHU S, et al. Retinal blood vessel segmentation based on Hessian matrix and level set[J]. Science Technology and Engineering, 2016, 16(10):44-49.)
[4] SOARES J V B, LEANDRO J J G, CESAR R M, Jr, et al. Retinal vessel segmentation using the 2-D Gabor wavelet and supervised classification[J]. IEEE Transactions on Medical Imaging, 2006, 25(9):1214-1221.
[5] 吴奎,蔡冬梅,贾鹏,等.基于2D Gabor小波与组合线检测算子的视网膜血管分割[J].科学技术与工程,2016,16(12):106-112.(WU K, CAI D M, JIA P, et al. Retinal vessel segmentation based on 2D Gabor wavelet and combined line operators[J]. Science Technology and Engineering, 2016, 16(12):106-112.)
[6] 吴奎.基于特征提取和监督学习的视网膜血管分割技术研究[D].太原:太原理工大学,2016:25-50.(WU K. Retinal vascular segmentation based on feature extraction and supervised learning[D]. Taiyuan:Taiyuan University of Technology, 2016:25-50.)
[7] ÀSLANI S, SARNEL H. A new supervised retinal vessel segmentation method based on robust hybrid features[J]. Biomedical Signal Processing and Control, 2016, 30:1-12.
[8] ZHU C Z, ZHOU B J, ZHAO R C, et al. Retinal vessel segmentation in colour fundus images using extreme learning machine[J]. Computer Medical Imaging and Graphics, 2017, 55:68-77.
[9] 朱承璋,崔锦恺,邹北骥,等.基于多特征融合和随机森林的视网膜血管分割[J].计算机辅助设计与图形学报,2017,29(4):584-592.(ZHU C Z, CUI J K, ZOU B J, et al. Retinal vessel segmentation based on multiple feature fusion and random forest[J]. Journal of Computer-Aided Design & Computer Graphics, 2017, 29(4):584-592.)
[10] FRANGI A F, NIESSEN W J, VINCKEN K L, et al. Multiscale vessel enhancement filtering[C]//MICCAI'98:Proceedings of the First International Conference on Medical Image Computing and Computer-Assisted Intervention, LNCS 1496. Berlin:Springer, 1998:130-137.
[11] AZZOPARDI G, STRISCIUGLIO N, VENTO M, et al. Trainable COSFIRE filters for vessel delineation with application to retinal images[J]. Medical Image Analysis, 2015, 19(1):46-57.
[12] 谭春林,曹鹏,黄华.基于相位一致和和模糊C均值的眼底血管分割[J].计算机工程与设计,2015,3(6):1551-1554.(TAN C L, CAO P, HUANG H. Vasculature segmentation of fundus image based on phase congruency and kernel fussy C-mean algorithm[J]. Computer Engineering and Design, 2015, 3(6):1551-155.)
[13] KOVESI P. Phase congruency:a low-level image invariant[J]. Psychology Research, 2000, 64(2):136-148.
[14] DASGUPTA A, SINGH S. A fully convolutional neural network based structured prediction approach towards the retinal vessel segmentation[C]//Proceedings of the 2017 International Symposium on Biomedical Imaging. Piscataway, NJ:IEEE, 2017:248-251.
[15] 徐光柱,张柳,邹耀斌,等.自适应脉冲耦合神经网络与匹配滤波器相结合的视网膜血管分割[J].光学精密工程,2017,25(3):756-764.(XU G Z, ZHANG L, ZOU Y B, et al. Retinal blood segmentation with adaptive PCNN and matched filter[J]. Optics and Precision Engineering, 2017, 25(3):756-764.) LI Yuanyuan, born in 1994, M. S. candidate. Her research interests include image processing.CAI Yiheng, born in 1974, Ph. D., associate professor. Her research interests include medical image processing.GAO Xurong, born in 1992, M. S. candidate. Her research interests include medical image processing.