融合空间聚类与结构特征的点集配准优化算法

doi:10.19665/j.issn1001-2400.20230411

摘要/Abstract

摘要：

在点集配准中,噪声、非刚性形变和误匹配的存在,产生了求解非线性最优空间变换困难的问题。针对这个问题引入局部约束条件,提出了一种采用局部空间聚类和邻域结构特征的点集配准优化算法(PR-SDCLS)。首先,利用点集空间距离矩阵构造运动一致性聚类子集和离群值聚类子集;然后,在运动一致性聚类子集中分别使用高斯混合模型拟合,并引入通过融合形状上下文特征描述子与加权空间距离获得考虑全局和局部特征的混合系数;最后,采用最大期望算法完成参数估计,实现了混合模型的非刚性点集配准模型;为了提高算法效率,模型变换采用再生核希尔伯特空间建模,并使用核近似策略。实验结果表明,该算法在涉及不同类型数据退化(变形、噪声、离群点、遮挡和旋转)的非刚性数据集上,面对大量异常值时具有良好的配准效果和鲁棒性,配准平均误差的均值在经典和先进的算法基础上降低了约42.053 8%。

关键词: 配准, 非刚性, 高斯混合模型, EM算法

Abstract:

The existence of noise,non-rigid deformation and mis-matching in point set registration results in the difficulty of solving nonlinear optimal space transformation.This paper introduces local constraints and proposes a point set registration optimization algorithm using spatial distance clustering and local structural features(PR-SDCLS).First,the motion consistency clustering subset and outlier clustering subset are constructed by using the point set space distance matrix;Then,the Gaussian mixture model is used to fit the motion consistency cluster subset,and the mixing coefficient considering global and local features is obtained by fusing the shape context feature descriptor and weighted spatial distance.Finally,the maximum expectation algorithm is used to complete the parameter estimation,and the non-rigid point set registration model of the Gaussian mixture model is realized.In order to improve the efficiency of the algorithm,the model transformation uses the reproducing kernel Hilbert space model,and uses the kernel approximation strategy.Experimental results show that the algorithm has a good registration effect and robustness in the face of a large number of outliers on non-rigid data sets involving different types of data degradation(deformation,noise,outliers,occlusion and rotation),and the mean value of registration average error is reduced by 42.053 8% on the basis of classic and advanced algorithms.

Key words: registration, non-rigid, Gaussian mixture model, EM algorithm

中图分类号:

TP391

胡欣,向迪源,秦皓,肖剑. 融合空间聚类与结构特征的点集配准优化算法[J]. 西安电子科技大学学报, 2023, 50(5): 95-106.

HU Xin,XIANG Diyuan,QIN Hao,XIAO Jian. Point set registration optimization algorithm using spatial clustering and structural features[J]. Journal of Xidian University, 2023, 50(5): 95-106.

图/表 9

图1

图2

图3

图4

表1

图5

表2

图6

表3

参考文献 26

[1]	马新科, 杨扬, 杨昆, 等. 基于模糊形状上下文与局部向量相似性约束的配准算法[J]. 自动化学报, 2020, 46(2):342-357.
	MA Xinke, YANG Yang, YANG Kun, et al. Registration Algorithm Based on Fuzzy Shape Context and Local Vector Similarity Constraint[J]. Acta Automatica Sinica, 2020, 46(2):342-357.
[2]	李佳炜, 江晶, 吴卫华, 等. 应用统计线性回归的系统误差最大似然配准[J]. 西安电子科技大学学报, 2021, 48(4):73-82.
	LI Jiawei, JIANG Jing, WU Weihua, et al. Maximum Likelihood Registration for Systemic Error Based on Statistical Linear Regression[J]. Journal of Xidian University, 2021, 48(4):73-82.
[3]	杨思燕, 曹文灿, 李世平. 结合高阶图模型与蚁群优化的图像匹配方法[J]. 西安电子科技大学学报, 2017, 44(1):159-164.
	YANG Siyan, CAO Wencan, LI Shiping. Second-Order Graph Model Ant and Colony Optimization Based Image Matching[J]. Journal of Xidian University, 2017, 44(1):159-164.
[4]	MA J, JIANG X, FAN A. Image Matching from Handcrafted to Deep Features:A Survey[J]. International Journal of Computer Vision, 2021, 129:23-79. doi: 10.1007/s11263-020-01359-2
[5]	何凯, 刘志国, 李大双, 等. 基于邻域结构和驱动力准则的非刚性点集配准[J]. 华南理工大学学报(自然科学版), 2022, 50(4):73-80.
	HE Kai, LIU Zhiguo, LI Dashuang, et al. Non-Rigid Point Set Registration Based on Neighborhood Structure and Driving Force Criterion[J]. Journal of South China University of Technology(Natural Science Edition), 2022, 50(4):73-80.
[6]	何淇淇, 林刚, 周杰, 等. 基于变分贝叶斯层次概率模型的非刚性点集配准[J]. 计算机学报, 2021, 44(9):1866-1887.
	HE Qiqi, LIN Gang, ZHOU Jie, et al. Non-Rigid Point Set Registration Based on Variational Bayes Hierarchical Probability Model[J]. Chinese Journal of Computers, 2021, 44(9):1866-1887.
[7]	TSIN Y, KANADE T. A Correlation-Based Approach to Robust Point Set Registration[C]//Proceedings of the European Conference on Computer Vision. Heidelberg:Springer, 2004:558-569.
[8]	GLAUNES J, TROUVÉ A, YOUNES L. Diffeomorphic Matching of Distributions:A New Approach for Unlabelled Point-Sets and Submanifolds Matching[C]//Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition. Piscataway:IEEE, 2004:712-718.
[9]	JIAN B, VEMURI B C. Robust Point Set Registration Using Gaussian Mixture Models[J]. IEEE Transactions on Pattern Analysis and Machine Intelligence, 2011, 33(8):1633-1645. doi: 10.1109/TPAMI.2010.223 pmid: 21173443
[10]	CAMPBELL D, PETERSSON L. An Adaptive Data Representation for Robust Point-Set Registration and Merging[C]//Proceedings of the IEEE International Conference on Computer Vision. Piscataway:IEEE, 2015:4292-4300.
[11]	MYRONENKO A, SONG X. Point Set Registration:Coherent Point Drift[J]. IEEE Transactions on Pattern Analysis and Machine Intelligence, 2010, 32(12):2262-2275. doi: 10.1109/TPAMI.2010.46
[12]	HORAUD R, FORBES F, YGUEL M, et al. Rigid and Articulated Point Registration with Expectation Conditional Maximization[J]. IEEE Transactions on Pattern Analysis and Machine Intelligence, 2011, 33(3):587-602. doi: 10.1109/TPAMI.2010.94 pmid: 20421669
[13]	ZHANG S, YANG K, YANG Y, et al. Non-Rigid Point Set Registration Using Dual-Feature Finite Mixture Model and Global-Local Structural Preservation[J]. Pattern Recognition, 2018, 80:183-195. doi: 10.1016/j.patcog.2018.03.004
[14]	ZHANG S, YANG Y, YANG K, et al. Point Set Registration with Global-Local Correspondence and Transformation Estimation[C]//Proceedings of the IEEE International Conference on Computer Vision. Piscataway:IEEE, 2017:2669-2677.
[15]	LAWIN F J, DANELLJAN M, KHAN F, et al. Density Adaptive Point Set Registration[C]//Proceedings of the IEEE International Conference on Computer Vision. Piscataway:IEEE, 2018:3829-3837.
[16]	JIAN B, VEMURI B. A Robust Algorithm for Point Set Registration Using Mixture of Gaussians[C]//Tenth IEEE International Conference on Computer Vision. Piscataway:IEEE, 2005:1246-1251.
[17]	MA J, ZHAO J, TIAN A, et al. Robust Point Matching via Vector Field Consensus[J]. IEEE Transactions on Image Processing, 2014, 23(4):1706-1721. doi: 10.1109/TIP.2014.2307478 pmid: 24808341
[18]	YANG Y, ONG S H, FOONG K W C. A Robust Global and Local Mixture Distance Based Non-Rigid Point Set Registration[J]. Pattern Recognition, 2015, 48(1):156-173. doi: 10.1016/j.patcog.2014.06.017
[19]	MA J, ZHAO J, YUILLE A L. Non-Rigid Point Set Registration by Preserving Global and Local Structures[J]. IEEE Transactions on Image Processing, 2016, 25 (1):53-64. doi: 10.1109/TIP.2015.2467217 pmid: 26276991
[20]	JIANG X, MA J, JIANG J, et al. Robust Feature Matching Using Spatial Clustering with Heavy Outliers[J]. IEEE Transactions on Image Processing, 2020, 29:736-746. doi: 10.1109/TIP.83
[21]	彭磊, 杨秀云, 张裕飞, 等. 基于全局与局部相似性测度的非刚性点集配准[J]. 计算机应用, 2019, 39(10):3028-3033. doi: 10.11772/j.issn.1001-9081.2019040681
	PENG Lei, YANG Xiuyun, ZHANG Yufei, et al. Non-Rigid Point Set Registration Based on Global and Local Similarity Measurement[J]. Journal of Computer Applications, 2019, 39(10):3028-3033. doi: 10.11772/j.issn.1001-9081.2019040681
[22]	HIROSE O. A Bayesian Formulation of Coherent Point Drift[J]. IEEE Transactions on Pattern Analysis and Machine Intelligence, 2021, 43(7):2269-2286. doi: 10.1109/TPAMI.2020.2971687
[23]	ZHU J, GUO R, LI Z, et al. Registration of Multi-View Point Sets Under the Perspective of Expectation-Maximization[J]. IEEE Transactions on Image Processing, 2020, 29:9176-9189. doi: 10.1109/TIP.83
[24]	WANG G, CHEN Y. SCM:Spatially Coherent Matching with Gaussian Field Learning for Nonrigid Point Set Registration[J]. IEEE Transactions on Neural Networks and Learning Systems, 2021, 32(1):203-213. doi: 10.1109/TNNLS.5962385
[25]	CHUI H, RANGARAJAN A. A New Point Matching Algorithm for Non-Rigid Registration[J]. Computer Vision and Image Understanding, 2003, 89(2-3):114-141. doi: 10.1016/S1077-3142(03)00009-2
[26]	ZHENG Y, DOERMANN D. Robust Point Matching for Nonrigid Shapes by Preserving Local Neighborhood Structures[J]. IEEE Transactions on Pattern Analysis and Machine Intelligence, 2006, 28(4):643-649. pmid: 16566512

退化类型	PR-SDCLS	PR-GLS	RFM-SCAN	VFC	CPD
旋转	0.425 836	1.018 762	1.388 470	1.349 078	1.017 443
形变	0.263 519	0.273 834	0.402 068	0.956 679	0.272 664
噪声	0.352 121	0.357 400	0.713 039	0.943 008	0.353 570
遮挡	0.272 809	0.427 621	0.537 245	0.964 909	0.393 586
离群点	0.203 299	0.289 748	0.298 733	0.946 830	0.288 850

退化类型	D123	D12	D13	D23
旋转	0.327 904	0.425 903	0.448 861	0.446 003
形变	0.258 162	0.263 840	0.264 654	0.264 391
噪声	0.170 481	0.352 526	0.319 863	0.321 300
遮挡	0.271 300	0.272 971	0.273 069	0.272 744
离群点	0.288 550	0.289 126	0.289 855	0.289 591

退化类型	系数1	系数2	系数3
旋转	0.476 864	0.437 650	0.407 940
形变	0.265 072	0.263 519	0.263 153
噪声	0.352 169	0.314 476	0.3114 56
遮挡	0.272 809	0.272 468	0.271 839
离群点	0.291 178	0.288 850	0.288 331