基于有向超图的图像匹配算法

详细信息    查看全文 | 推荐本文 |

英文篇名：Image Matching Algorithm Based on Directed Hypergraph 作者：朱明 ; 张健 ; 梁栋 ; 唐俊 ; 张艳 英文作者：ZHU Ming;ZHANG Jian;LIANG Dong;TANG Jun;ZHANG Yan;National Engineering Research Center for Agro-Ecological Big Data Analysis and Application, Anhui University;School of Electronics and Information Engineering, Anhui University;Key Laboratory of Polarization Imaging Detection Technology in Anhui Province; 关键词：超图 ; 三元组 ; 邻接张量 ; 凸凹松弛 ; 匹配 英文关键词：hypergraph;;triples;;adjacency tensor;;convex-concave relaxation;;matching 中文刊名：HNLG 英文刊名：Journal of South China University of Technology(Natural Science Edition) 机构：安徽大学农业生态大数据分析与应用技术国家地方联合工程研究中心;安徽大学电子信息工程学院;偏振光成像探测技术安徽省重点实验室; 出版日期：2019-02-15 出版单位：华南理工大学学报(自然科学版) 年：2019 期：v.47;No.389 基金：国家自然科学基金资助项目(61501003,61772032,61672032);; 偏振光成像探测技术安徽省重点实验室开放课题(2016-KFJJ-002)~~ 语种：中文; 页：HNLG201902007 页数：9 CN：02 ISSN：44-1251/T 分类号：56-64

摘要

为了提高复杂变换下图模型的匹配精度,提出了一种基于有向超图的图像匹配算法.该算法首先分别在两个待匹配的特征点集中构造3一致超图,计算每条超边所包含三元组的权值,然后利用这些权值来构造加权邻接张量,最后通过凸凹松弛算法实现图像匹配.模拟和真实图像的实验结果表明,文中算法能够获得更高的匹配精度,对于复杂变换的图像也有很好的匹配效果.
        An algorithm for images matching based on directed hypergraph was proposed to improve the matching accuracy of graph model under the complex transformation. Firstly, 3-uniform hypergraphs in the two feature points set to be matched was constructed by the algorithm, and the weights of the triples contained in each hyperedge was calculated, then these weights were used to construct weighted adjacency tensors.Finally, image matching was achieved by convex-concave relaxation algorithm. The experiments with both simulate and real results show that the proposed algorithm can achieve higher matching accuracy and good matching effect for complex transformed images.

引文

[1] 梁栋,朱明,唐俊,等.基于局部相对形状上下文与Q-谱的点模式匹配算法 [J].电子学报,2012,40(4):636- 641.LIANG Dong,ZHU Ming,TANG Jun,et al.A point pattern matching algorithm based on local relative shape context and Q-spectra [J].Acta Electronica Sinica,2012,40(4):636- 641.
    [2] TANG Jun,SHAO Lin,ZHEN Xian-tong.Robust point pattern matching based on spectral context [J].Pattern Recognition,2014,47(3):1469- 1484.
    [3] SUH Y,ADAMCZEWSKI K,LEE K M.Subgraph mat-ching using compactness prior for robust feature corres-pondence [C]//Proceedings of 2015 IEEE Conference on Computer Vision and Pattern Recognition.Boston:IEEE,2015:5070- 5078.
    [4] JIANG Bo,TANG Jin,CAO Xiaochun,et al.Lagrangian relaxation graph matching [J].Pattern Recognition,2017,61:255- 265.
    [5] ZASS R,SHASHUA A.Probabilistic graph and hyper-graph matching [C]//Proceedings of 2008 IEEE Confe-rence on Computer Vision and Pattern Recognition.Anchorage:IEEE,2008:1221- 1228.
    [6] DUCHENNE O,BACH F,KWEON I,et al.A tensor-based algorithm for high-order graph matching [C]//Proceedings of 2009 IEEE Computer Society Conference on Computer Vision and Pattern Recognition Workshops.Miami:IEEE,2009:1980- 1987.
    [7] LEE J,CHO M,LEE K M.Hyper-graph matching via reweighted random walks [C]//Proceedings of 2011 IEEE Conference on Computer Vision and Pattern Recognition.Washington:IEEE,2011:1633- 1640.
    [8] YANG Xu,QIAO Hong,LIU Zhi-Yong.Point correspondence by a new third order graph matching algorithm [J].Pattern Recognition,2016,65:108- 118.
    [9] YANG Xu,LIU Zhi-Yong,QIAO Hong,et al.Probabilistic hypergraph matching based on affinity tensor updating [J].Neurocomputing,2017,269(20):142- 147.
    [10] LEORDEANU M,HEBERT M.A spectral technique for correspondence problems using pairwise constraints [C]//Proceedings of the 10th IEEE International Conference on Computer Vision.Beijing:IEEE,2005:1482- 1489.
    [11] CHO M,LEE J,LEE K M.Reweighted random walks for graph matching [C]//Proceedings of 2010 European Conference on Computer Vision.Heraklion:Springer,2010:492- 505.
    [12] PARK S,PARK S K,HEBERT M.Fast and scalable approximate spectral matching for higher order graph matching [J].IEEE Transactions on Pattern Analysis and Machine Intelligence,2014,36(3):479- 492.
    [13] LIU Zhi-Yong,QIAO Hong.GNCCP:graduated non-convexity and concavity procedure [J].IEEE Transactions on Pattern Analysis and Machine Intelligence,2014,36(6):1258- 1267.
    [14] LIU Zhi-Yong,QIAO Hong,YANG Xu,et al.Graph mat-ching by simplified convex-concave relaxation procedure [J].International Journal of Computer Vision,2014,109(3):169- 186.
    [15] ZASLAVSKIY M,BACH F,VERT J P.A path following algorithm for the graph matching problem [J].IEEE Transactions on Pattern Analysis and Machine Intelligence,2009,31(12):2227- 2242.
    [16] YANG Xu,QIAO Hong,LIU Zhi-Yong.Feature correspondence based on directed structural model matching [J].Image & Vision Computing,2015,33:57- 67.
    [17] KUHN D,OSTHUS D,TREGLOWN A.Matchings in 3-uniform hypergraphs [J].Journal of Combinatorial Theory(Series B),2013,103(2):291- 305.
    [18] UMEYAMA S.An eigendecomposition approach to wei-ghted graph matching problems [J].IEEE Transactions on Pattern Analysis & Machine Intelligence,1988,10(5):695- 703.
    [19] 鲍文霞,余国芬,胡根生,等.基于马氏距离谱特征的图像匹配算法 [J].华南理工大学学报(自然科学版),2017,45(10):114- 120.BAO Wen-xia,YU Guo-fen,HU Gen-sheng,et al.Image matching algorithm based on mahalanobis-distance spectral features [J].Journal of South China University of Technology(Natural Science Edition),2017,45(10):114- 120.
    [20] CHUI H,RANGARJAN A.A new point matching algorithm for non-rigid registration [J].Computer Vision and Image Understanding,2003,89(2):114- 141.