Correspondence Calculation of Non-isometric 3D Point Shapes Based on Smooth Attention and Spectral Up-sampling Refinement
-
摘要: 为了解决非等距3维点云模型对应关系计算易受模型大尺度形变影响而导致对应失真、准确率低且平滑性差的问题，该文提出一种结合平滑注意力与谱上采样细化的非等距3维点云模型对应关系计算新方法。首先，利用点所在表面的几何特征信息设计平滑注意力机制与平滑感知模块，提高特征对大尺度形变区域非刚性变换的感知能力；其次，将深度函数映射模块与平滑正则化约束相结合，提升函数映射计算结果的平滑性；最后，在谱上采样细化模块中，以多分辨率重建的方式得到最终的逐点映射结果。实验结果表明，与已有算法相比，本算法在FAUST、SCAPE和SMAL数据集上构建的对应关系测地误差最小，处理大尺度形变模型时，能够提升逐点映射的平滑性和全局准确率。Abstract: To address the problem that the correspondence calculation of non-isometric 3D point cloud shape is easily affected by large-scale distortions, which often leads to corresponding distortions, low accuracy, and poor smoothness, a new algorithm of shape correspondence calculation for non-isometric 3D point cloud is proposed, which combines smooth attention with spectral up-sampling refinement. Firstly, a smooth attention mechanism and a smooth perception module are designed using the geometric feature information of the surface on which the points are located to improve the perception ability of the features for non-rigid transformations in large-scale deformation areas. Secondly, the deep functional maps module is combined with smooth regularization constraints to improve the smoothness of the functional maps calculation results. Finally, the final point-by-point mapping result is obtained using a multi-resolution reconstruction method in the spectral up-sampling refinement module. Experimental results show that the proposed algorithm has the smallest geodesic error in the correspondence constructed on the FAUST, SCAPE, and SMAL datasets compared with existing algorithms. It can improve the smoothness and global accuracy of point-by-point mapping for shapes with large-scale deformation.
-
[1] DENG Bailin, YAO Yuxin, DYKE R M, et al. A survey of non-rigid 3D registration[J]. Computer Graphics Forum, 2022, 41(2): 559–589. doi: 10.1111/cgf.14502. [2] OVSJANIKOV M, BEN-CHEN M, SOLOMON J, et al. Functional maps: A flexible representation of maps between shapes[J]. ACM Transactions on Graphics (TOG), 2012, 31(4): 30. doi: 10.1145/2185520.2185526. [3] WU Yan, YANG Jun, and ZHAO Jinlong. Partial 3D shape functional correspondence via fully spectral eigenvalue alignment and upsampling refinement[J]. Computers & Graphics, 2020, 92: 99–113. doi: 10.1016/j.cag.2020.09.004. [4] LITANY O, REMEZ T, RODOLÀ E, et al. Deep functional maps: Structured prediction for dense shape correspondence[C]. The IEEE International Conference on Computer Vision, Venice, Italy, 2017: 5660–5668. doi: 10.1109/ICCV.2017.603. [5] TOMBARI F, SALTI S, and DI STEFANO L. Unique signatures of histograms for local surface description[C]. The 11th European Conference on Computer Vision, Heraklion, Greece, 2010: 356–369. doi: 10.1007/978-3-642-15558-1_26. [6] ROUFOSSE J M, SHARMA A, and OVSJANIKOV M. Unsupervised deep learning for structured shape matching[C]. The IEEE/CVF International Conference on Computer Vision, Seoul, Korea (South), 2019: 1617–1627. doi: 10.1109/ICCV.2019.00170. [7] MARIN R, RAKOTOSAONA M J, MELZI S, et al. Correspondence learning via linearly-invariant embedding[C]. Proceedings of the 34th International Conference on Neural Information Processing Systems, Vancouver, Canada, 2020: 136. [8] ATTAIKI S, PAI G, and OVSJANIKOV M. DPFM: Deep partial functional maps[C]. The 2021 International Conference on 3D Vision, London, United Kingdom, 2021: 175–185. doi: 10.1109/3DV53792.2021.00040. [9] DONATI N, CORMAN E, and OVSJANIKOV M. Deep orientation-aware functional maps: Tackling symmetry issues in shape matching[C]. The IEEE/CVF Conference on Computer Vision and Pattern Recognition, New Orleans, USA, 2022: 732–741. doi: 10.1109/CVPR52688.2022.00082. [10] 管焱然, 奥利弗·范凯克, 管有庆. 基于重心映射的三角形网格参数化方法研究与实现[J]. 北京邮电大学学报, 2019, 42(5): 83–90. doi: 10.13190/j.jbupt.2018-266.GUAN Yanran, VAN KAICK O, and GUAN Youqing. Research and implementation of triangle mesh parameterization method based on barycentric mapping[J]. Journal of Beijing University of Posts and Telecommunications, 2019, 42(5): 83–90. doi: 10.13190/j.jbupt.2018-266. [11] DONATI N, SHARMA A, and OVSJANIKOV M. Deep geometric functional maps: Robust feature learning for shape correspondence[C]. The IEEE/CVF Conference on Computer Vision and Pattern Recognition, Seattle, USA, 2020: 8589–8598. doi: 10.1109/CVPR42600.2020.00862. [12] EYNARD D, RODOLÀ E, GLASHOFF K, et al. Coupled functional maps[C]. The 2016 Fourth International Conference on 3D Vision, Stanford, USA, 2016: 399–407. doi: 10.1109/3DV.2016.49. [13] PINKALL U and POLTHIER K. Computing discrete minimal surfaces and their conjugates[J]. Experimental Mathematics, 1993, 2(1): 15–36. doi: 10.1080/10586458.1993.10504266. [14] VASWANI A, SHAZEER N, PARMAR N, et al. Attention is all you need[C]. Proceedings of the 31st International Conference on Neural Information Processing Systems, Long Beach, USA, 2017: 6000–6010. [15] SHARP N, ATTAIKI S, CRANE K, et al. DiffusionNet: Discretization agnostic learning on surfaces[J]. ACM Transactions on Graphics (TOG), 2022, 41(3): 27. doi: 10.1145/3507905. [16] MELZI S, REN Jing, RODOLÀ E, et al. ZoomOut: Spectral upsampling for efficient shape correspondence[J]. ACM Transactions on Graphics (TOG), 2019, 38(6): 155. doi: 10.1145/3355089.3356524. [17] BOGO F, ROMERO J, LOPER M, et al. FAUST: Dataset and evaluation for 3D mesh registration[C]. The IEEE Conference on Computer Vision and Pattern Recognition, Columbus, USA, 2014: 3794–3801. doi: 10.1109/CVPR.2014.491. [18] ANGUELOV D, SRINIVASAN P, KOLLER D, et al. SCAPE: Shape completion and animation of people[C]. ACM SIGGRAPH 2005 Papers, Los Angeles, USA, 2005: 408–416. doi: 10.1145/1186822.1073207. [19] ZUFFI S, KANAZAWA A, JACOBS D W, et al. 3D menagerie: Modeling the 3D shape and pose of animals[C]. The IEEE Conference on Computer Vision and Pattern Recognition, Honolulu, USA, 2017: 5524–5532. doi: 10.1109/CVPR.2017.586. -
下载:
图(9)
计量
- 文章访问数: 59
- HTML全文浏览量: 17
- PDF下载量: 9
- 被引次数: 0
下载:
