基于超边图匹配的视网膜眼底图像配准算法

删除或更新信息，请邮件至freekaoyan#163.com(#换成@)

清华大学辅仁网/2017-07-07

基于超边图匹配的视网膜眼底图像配准算法

邓可欣(

)

Retinal image registration based on hyper-edge graph matching

Kexin DENG(

)

School of Electronic Engineering, Xidian University, Xi'an 710071, China

摘要:

HTML
输出: BibTeX | EndNote (RIS) 背景资料

文章导读

摘要视网膜眼底图像配准是临床眼科疾病诊断和治疗中的一个关键环节。针对眼底图像配准过程中大范围视场变化和过分割结构噪声等问题,该文提出了一种改进的基于图的视网膜图像血管匹配方法。将血管交叉点表示成图的顶点,把特征点间沿血管路径的相邻关系表示成边,进而在视网膜血管结构图中构造路径超边来刻画更高阶多元特征关系。在此基础上,实现了一种全自动的视网膜眼底图像配准算法。包括: 第一步,通过多尺度Gabor滤波算法来检测和提取视网膜血管网络; 第二步,利用一种高效的谱松弛匹配算法来求解两个路径超边图的顶点匹配对应关系。最后,通过特征匹配召回率统计和配准的血管中线距离误差两方面的实验,证明该文提出算法是有效和准确的。

关键词 ：图匹配,超边,谱松弛匹配,血管检测,视网膜图像配准

Abstract：Retinal fundus image registration is a key step in clinical eye disease diagnosis and treatment. Targeting the difficulties in dealing with pair of images with large field of view differences or with great structural noises caused by over-segmentation, this paper presents an improved graph-based algorithm to match retinal vessel networks by utilizing higher-order relations constructed from the vessel structural graphs whose nodes represent vascular bifurcations with the edges describing relations between feature points. The method performs in a fully automatic fashion, with a multi-scale Gabor filter first employed for detection and extraction of retinal vessels and with the correspondences then recovered between two pathwise hyper-edge graphs using an efficient pairwise spectral matching scheme. Experimental evaluation shows that the developed method is effective and accurate in terms of the feature recall rate and the vascular CEM distance.

Key words：graph matching hyper-edge spectral matching vessel detection retinal image registration

收稿日期: 2014-03-31 出版日期: 2015-09-03

ZTFLH:

基金资助:

引用本文:

邓可欣. 基于超边图匹配的视网膜眼底图像配准算法[J]. 清华大学学报（自然科学版）, 2014, 54(5): 568-574.
Kexin DENG. Retinal image registration based on hyper-edge graph matching. Journal of Tsinghua University(Science and Technology), 2014, 54(5): 568-574.

链接本文:

http://jst.tsinghuajournals.com/CN/或 http://jst.tsinghuajournals.com/CN/Y2014/V54/I5/568

图表:

在较强分割噪声影响下的血管结构图模型匹配结果

不同尺度下Gabor血管检测结果图像

多分辨率血管检测分割结果

超边图模型及其二阶简单图近似

视网膜图像血管路径超边图构造与匹配图示

测试样本序列	噪声率	匹配召回率
测试样本序列	噪声率	GA	BGA	SM	HE-SM
R01-02	0.45	0.43	0.54	0.49	0.54
R03-04	0.42	0.67	0.58	0.23	0.82
R05-06	0.28	0.44	0.68	0.73	0.76
R07-08	0.32	0.26	0.61	0.79	0.65
R09-10	0.23	0.42	0.75	0.83	0.86
R11-12	0.55	0.13	0.30	0	0.70
R13-14	0.49	0.26	0.55	0.53	0.77
R15-16	0.53	0.11	0.33	0.44	0.81
均值	0.41	0.34	0.54	0.51	0.74

视网膜血管结构图匹配算法对比实验

视网膜眼底图像配准示例

测试样本序列	图像交叠率/%	血管中线距离/像素
测试样本序列	图像交叠率/%	GM-ICP	HE-GM-ICP	GDB-ICP
R01-02	63.02	0.68	0.68	0.71
R03-04	64.72	0.79	0.77	0.74
R05-06	85.49	0.74	0.74	0.73
R07-08	80.98	1.05	1.05	1.02
R09-10	83.00	0.66	0.65	0.64
R11-12	51.02	0.96	0.89	0.85
R13-14	51.62	1.22	1.14	1.03
R15-16	63.17	0.86	0.81	0.80

视网膜眼底图像配准对比实验

参考文献:

[1]	DENG Kexin, TIAN Jie, ZHENG Jian, et al.Retinal fundus image registration via vascular structure graph matching[J]. International Journal of Biomedical Imaging, 2010, doi: doi: 10.1155/2010/906067
[2]	Legg P, Rosin P, Marshall D, et al.Improving accuracy and efficiency of mutual information for multi-modal retinal image registration using adaptive probability density estimation[J]. Computerized Medical Imaging and Graphics, 2013, 37(7-8): 597-606.
[3]	TSAI Chialing, LI Chunyi, YANG Gehua, et al.The edge-driven dual-bootstrap iterative closest point algorithm for registration of multimodal fluorescein angiogram sequence[J]. IEEE Transactions on Medical Imaging, 2010, 29(3): 636-649.
[4]	Stewart C, TSAI Chialing, Roysam B. The dual-bootstrap iterative closest point algorithm with application to retinal image registration[J]. IEEE Transactions on Medical Imaging, 2003, 22(11): 1379-1394.
[5]	ZHENG Yuanjie, Daniel E, Allan A, et al.Landmark matching based retinal image alignment by enforcing sparsity in correspondence matrix[J]. Medical Image Analysis, 2013, doi: doi: 10.1016/j.media.2013.09.009
[6]	Perez R A, Cabido R, Trucco E, et al.RERBEE: Robust efficient registration via bifurcations and elongated elements applied to retinal fluorescein angiogram sequences[J]. IEEE Transactions on Medical Imaging, 2012, 31(1): 140-150.
[7]	Dashtbozorg B, Mendonca A M, Campilho A. An automatic graph-based approach for artery/vein classification in retinal images[J]. IEEE Transactions on Image Processing, 2014, 23(3): 1073-1083.
[8]	XU Xiayu, Meindert N, QI Song, et al.Vessel boundary delineation on fundus images using graph-based approach[J]. IEEE Transactions on Medical Imaging, 2011, 30(6): 1184-1191.
[9]	Serradell E, Przemyslaw G, Jan K, et al.Robust non-rigid registration of 2d and 3d graphs [C]// IEEE Conference on Computer Vision and Pattern Recognition. IEEE, 2012: 996-1003.
[10]	Soares J, Leandro J, Cesar R, et al.Retinal vessel segmentation using the 2d gabor wavelet and supervised classification[J]. IEEE Transactions on Medical Imaging, 2006, 25(9): 1214-1222.
[11]	Gold S, Rangarajan A. A graduated assignment algorithm for graph matching[J]. IEEE Transactions on Pattern Analysis and Machine Intelligence, 1996, 18(4): 377-388.
[12]	Leordeanu M, Hebert M. A spectral technique for correspondence problems using pairwise constraints [C]// IEEE International Conference on Computer Vision. IEEE, 2005: 1482-1489.
[13]	Cour T, Srinivasan P, SHI Jianbo. Balanced graph matching [C]// NIPS: Neural Information Processing Systems Foundation. 2006: 313-320.
[14]	Schellewald C, Schnörr C. Probabilistic subgraph matching based on convex relaxation [C]// Energy Minimization Methods in Computer Vision and Pattern Recognition. Springer, 2005: 171-186.
[15]	Duchenne O, Bach F, Kweon I, et al.A tensor-based algorithm for high-order graph matching[J]. IEEE Transactions on Pattern Analysis and Machine Intelligence, 2011, 33(12): 2383-2395.
[16]	Zass R, Shashua A. Probabilistic graph and hypergraph matching [C]// IEEE Conference on Computer Vision and Pattern Recognition. IEEE, 2008: 1-8.
[17]	Kang U, Hebert M, Park S. Fast and scalable approximate spectral graph matching for correspondence problems [C]// Information Sciences 220. Online Fuzzy Machine Learning and Data Mining. 2013: 306-318.
[18]	CHUI Haili, Rangarajan A. A new point matching algorithm for non-rigid registration[J]. Computer Vision and Image Understanding, 2003, 89(2): 114-141.

[1]	王振波, 张君, 罗孙一鸣. 喷水法成型纤维网增强水泥基板材抗弯性能[J]. 清华大学学报（自然科学版）, 2014, 54(5): 551-555.
[2]	徐悟, 于清, 尧国皇. 初应力对钢管混凝土叠合柱轴压性能影响[J]. 清华大学学报（自然科学版）, 2014, 54(5): 556-562.
[3]	魏亚, 姚湘杰. 约束状态下混凝土拉伸徐变模型[J]. 清华大学学报（自然科学版）, 2014, 54(5): 563-567.
[4]	罗逍, 姚远, 张金换. 一种毫米波雷达和摄像头联合标定方法[J]. 清华大学学报（自然科学版）, 2014, 54(3): 289-293.
[5]	朱兵, 陈洵欢, 张文俊, 胡山鹰, 金涌. 中国合成氨行业清洁生产潜力分析[J]. 清华大学学报（自然科学版）, 2014, 54(3): 309-313.
[6]	陆瑶, 袁敏峰. 新股“破发”与IPO定价效率: 基于股东特征的分析[J]. 清华大学学报（自然科学版）, 2014, 54(3): 381-393.
[7]	张金换, 刘卫国, 李景涛, 赵福全. 行人头型冲击器试验有限元建模及敏感参数分析[J]. 清华大学学报（自然科学版）, 2014, 54(3): 294-298.
[8]	许述财, 邹猛, 魏灿刚, 王龙臻, 张金换, 于用军. 仿竹结构薄壁管的轴向耐撞性分析及优化[J]. 清华大学学报（自然科学版）, 2014, 54(3): 299-304.
[9]	屈岳波, 曹彬, 王梁, 张伯奇, 吴健栋, 蔡志鹏, 潘际銮. 窄间隙埋弧焊接头熔合区弱化的研究[J]. 清华大学学报（自然科学版）, 2014, 54(3): 305-308.
[10]	林源, 梁舒, 王生进. 基于非刚性ICP的三维人脸数据配准算法[J]. 清华大学学报（自然科学版）, 2014, 54(3): 334-340.
[11]	张珂, 丁巧林, 刘涛, 赵伟. 基于细节空间关系的自然语言组合描述方法[J]. 清华大学学报（自然科学版）, 2014, 54(3): 341-347.
[12]	孙轶轩, 邵春福, 计寻, 朱亮. 基于ARIMA与信息粒化SVR组合模型的交通事故时序预测[J]. 清华大学学报（自然科学版）, 2014, 54(3): 348-353.
[13]	强彦, 裴博, 赵涓涓, 路景贵. 模糊支持向量机在肺结节良恶性分类中的应用[J]. 清华大学学报（自然科学版）, 2014, 54(3): 354-359.
[14]	朱强, 孙玉强. 一种基于信任度的协同过滤推荐方法[J]. 清华大学学报（自然科学版）, 2014, 54(3): 360-365.
[15]	丁荣涛. 基于合作博弈的港口物流链云服务组织方法[J]. 清华大学学报（自然科学版）, 2014, 54(3): 366-372.