北京生物医学工程

基于SHOT与目标函数对称ICP的低重叠率术前术中点云配准算法

Low overlap rate preoperative and intraoperative point cloud registration algorithm based on SHOT and symmetric objective function ICP

作者：严文磬张立静王玮武博

单位：首都医科大学生物医学工程学院（北京 100069）<br />首都医科大学临床生物力学应用基础研究北京市重点实验室（北京 100069)<br />首都医科大学宣武医院（北京 100053)<br />通信作者：武博，副教授。E-mail：wubogo@ccmu.edu.cn

关键词：点云配准；SHOT；目标函数对称ICP；低重叠率

分类号：R318.04

出版年·卷·期（页码）：2023·42·2（111-116）

摘要：

目的配准是术前影像引导的椎弓根螺钉内固定术中的重要环节。术前CT影像三维重建获得的点云和术中捕获的暴露部位点云重叠率低，易受噪声、遮挡等因素的干扰，使点云配准更具挑战性。本文采用局部特征和距离度量相结合的方式应对低重叠率配准问题。方法首先利用方向直方图描述子(signature of histograms of orientations, SHOT)描述子和随机抽样一致算法(random sample consensus, RANSAC)提取并匹配几何特征相似的点，完成初始对齐。应用目标函数对称ICP，通过最小化对称化点到面目标函数得到最终变换矩阵。对来源于SpineWeb公开数据集的5组腰椎数据进行配准实验。结果术前术中点云配准实验中平均配准误差为0.128 mm，平均运行时间为5.750 s。结论实验结果验证了该算法在低重叠率术前术中点云配准中的有效性，使得医生能及时根据配准结果调整手术器械，从而提高椎弓根螺钉置入准确率。

Objective Registration is an important part of preoperative image-guided pedicle screw fixation. The point cloud obtained by the 3D reconstruction of the preoperative CT image and the point cloud of the exposed part captured during the operation have a low overlap rate and are easily interfered by factors such as noise and occlusion, which makes the point cloud registration more challenging. This paper adopted the combination of local geometric features and distance measurement to deal with the problem of low overlap registration. Methods Firstly, the points with similar geometric features were extracted and matched using the signature of histograms of orientations descriptor(SHOT) and the random sample consensus(RANSAC) to complete the initial alignment. Applying the objective function symmetric ICP algorithm, the final transformation matrix was obtained by minimizing the symmetric point-to-surface objective function. Registration experiments were performed on five groups of lumbar spine data from the SpineWeb public dataset. Results In the preoperative and intraoperative point cloud registration experiments, the average registration error was 0.128 mm, and the average running time was 5.750 s. Conclusions The experimental results verify the effectiveness of the algorithm in the preoperative and intraoperative point cloud registration with low overlap rate, which enables doctors to adjust the surgical instruments according to the registration results in time, thus improving the accuracy of pedicle screw placement.

参考文献：

[1] 刘政, 李宏伟, 王爽. 胸腰椎椎弓根螺钉置入技术的比较分析与展望[J]. 中国组织工程研究, 2017, 21(19): 3102-3107.
Liu Z, Li HW, Wang S. Pedicle screw fixation for thoracic and lumbar diseases: a comparative analysis and prospects[J]. Chinese Journal of Tissue Engineering Research, 2017, 21(19): 3102-3107.
[2] Swiatek PR, McCarthy MH, Weiner J, et al. Intraoperative image guidance for lateral position surgery[J]. Annals of Translational Medicine, 2021, 9(1): 90.
[3] Li SJ, Yan WP, Chen ML, et al. Virtual bronchoscopic navigation without fluoroscopy guidance for peripheral pulmonary lesions in inexperienced pulmonologist[J]. Chinese Journal of Cancer Research, 2020, 32(4): 530-539.
[4] Cai YL, Wu SJ, Fan XY, et al. A level-wise spine registration framework to account for large pose changes[J]. International Journal of Computer Assisted Radiology and Surgery, 2021, 16(6): 943-953.
[5] Shetty P, Yeole U, Singh V, et al. Navigated ultrasound-based image guidance during resection of gliomas: practical utility in intraoperative decision-making and outcomes[J]. Neurosurgical Focus, 2021, 50(1): E14.
[6] Zhao XZ, Xu WH, Tang ZH, et al. Accuracy of computer-guided implant surgery by a CAD/CAM and laser scanning technique [J]. The Chinese Journal of Dental Research: the Official Journal of the Scientific Section of the Chinese Stomatological Association (CSA), 2014, 17(1): 31-36.
[7] Birnbaum FA, Wang A, Brady CJ. Stereoscopic surgical recording using GoPro cameras: a low-cost means for capturing external eye surgery[J]. JAMA Ophthalmology, 2015, 133(12): 1483-1484.
[8] Li P, Wang RS, Wang YX, et al. Evaluation of the ICP algorithm in 3D point cloud registration[J]. IEEE Access, 2020, 8: 68030-68048.
[9] Aiger D, Mitra NJ, Cohen-Or D. 4-points congruent sets for robust pairwise surface registration[J]. ACM Transactions on Graphics, 2008, 27(3): 85.
[10] Mellado N, Aiger D, Mitra N. Super 4PCS: fast global point cloud registration via smart indexing[J]. Computer Graphics Forum, 2014, 33(5): 205-215.
[11] Raposo C, Barreto JP. Using 2 point+normal sets for fast registration of point clouds with small overlap[C]//2017 IEEE International Conference on Robotics and Automation. Singapore: IEEE Press, 2017: 5652-5658.
[12] Chetverikov D, Stepanov D, Krsek P. Robust Euclidean alignment of 3D point sets: the trimmed iterative closest point algorithm[J]. Image and Vision Computing, 2005, 23(3): 299-309.
[13] Yang JL, Li HD, Campbell D, et al. Go-ICP: a globally optimal solution to 3D ICP point-set registration[J]. IEEE Transactions on Pattern Analysis and Machine Intelligence, 2016, 38(11): 2241-2254.
[14] Rusu RB, Bradski G, Thibaux R, et al. Fast 3D recognition and pose using the viewpoint feature histogram[C]//2010 IEEE/RSJ International Conference on Intelligent Robots and Systems. Taipei, Taiwan, China: IEEE Press, 2010: 2155-2162.
[15] Rusu RB, Blodow N, Beetz M. Fast point feature histograms (FPFH) for 3D registration[C]//2009 IEEE International Conference on Robotics and Automation. Kobe, Japan: IEEE Press, 2009: 3212-3217.
[16] Salti S, Tombari F, Stefano LD. SHOT: unique signatures of histograms for surface and texture description[J]. Computer Vision and Image Understanding, 2014, 125: 251-264.
[17] Guo YL, Sohel F, Bennamoun M, et al. Rotational projection statistics for 3D local surface description and object recognition[J]. International Journal of Computer Vision, 2013, 105(1): 63-86.
[18] Shen ZQ, Ma X, Li YB. A hybrid 3D descriptor with global structural frames and local signatures of histograms[J]. IEEE Access, 2018, 6: 39261-39272.
[19] Wang X, Zhu XH, Ying SH, et al. An accelerated and robust partial registration algorithm for point clouds[J]. IEEE Access, 2020, 8: 156504-156518.
[20] Xiong FG, Dong B, Huo W, et al. A local feature descriptor based on rotational volume for pairwise registration of point clouds[J]. IEEE Access, 2020, 8: 100120-100134.
[21] Besl PJ, McKay ND. A method for registration of 3-D shapes[J]. IEEE Transactions on Pattern Analysis and Machine Intelligence, 1992, 14(2): 239-256.
[22] Chen Y, Medioni G. Object modelling by registration of multiple range images[J]. Image and Vision Computing, 1992, 10(3): 145-155.
[23] Segal A, Haehnel D, Thrun S. Generalized-ICP[C]//Robotics: Science and Systems. Seattle, WA, USA: MIT Press, 2009: 435.
[24] Wang FP, Xiao J, Wang Y, et al. An improved ICP method using Gaussian curvature[J]. Journal of University of Chinese Academy of Sciences, 2019, 36(5): 702-708.
[25] Geng N, Ma FF, Yang HJ, et al. Neighboring constraint-based pairwise point cloud registration algorithm[J]. Multimedia Tools and Applications, 2016, 75(24): 16763-16780.
[26] Koide K, Yokozuka M, Oishi S, et al. Voxelized GICP for fast and accurate 3D point cloud registration[C]//2021 IEEE International Conference on Robotics and Automation. Xi’an, China: IEEE Press, 2021: 11054-11059.
[27] Rusinkiewicz S. A symmetric objective function for ICP[J]. ACM Transactions on Graphics, 2019, 38(4): 1-7.?
[28] Huang XS, Zhang J, Wu Q, et al. A coarse-to-fine algorithm for matching and registration in 3D cross-source point clouds[J]. IEEE Transactions on Circuits and Systems for Video Technology, 2018, 28(10): 2965-2977.
[29] Fischler MA, Bolles RC. Random sample consensus: a paradigm for model fitting with applications to image analysis and automated cartography[J]. Communications of the ACM, 1981, 24(6): 381-395.
[30] Prokop M, Shaikh SA, Kim KS. Low overlapping point cloud registration using line features detection[J]. Remote Sensing, 2019, 12(1): 61.

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