Abstract
Structure plays a crucial role in point cloud completion. While many efforts have been made to recover geometric details of the target shape, it is non-trivial to recover global structures, especially when large areas are missing in the input partial point cloud. In this paper, we propose a novel point cloud completion approach named SAPCNet to reconstruct complete 3D shapes in a global structure-aware manner. To establish long-range dependencies within the global scope and efficiently capture the object structure, we propose a structure-aware enhancement module by introducing global attention to the seed point generation process. To explicitly enforce the category-specific structure recovery, we design a classification-supervised training strategy to enable the network to generate structures more consistent with object category properties. Moreover, we design a category-guided adaptive seed selection module to generate seed points in the first stage. A redundant structure elimination loss is also introduced to reduce the ill-posed generation of heavily missing regions in the incomplete point cloud. Qualitative and quantitative evaluations on the PCN dataset demonstrate that our method outperforms state-of-the-art completion approaches.
This work was supported by the National Natural Science Foundation of China under Grant 62076230.
This is a preview of subscription content, log in via an institution to check access.
Access this chapter
Tax calculation will be finalised at checkout
Purchases are for personal use only
Similar content being viewed by others
References
Chang, A.X., et al.: ShapeNet: an information-rich 3D model repository. arXiv preprint arXiv:1512.03012 (2015)
Groueix, T., Fisher, M., Kim, V.G., Russell, B.C., Aubry, M.: A papier-mâché approach to learning 3D surface generation. In: Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, pp. 216–224 (2018)
Guo, M.H., Cai, J.X., Liu, Z.N., Mu, T.J., Martin, R.R., Hu, S.M.: PCT: point cloud transformer. Comput. Visual Media 7, 187–199 (2021)
Huang, Z., Yu, Y., Xu, J., Ni, F., Le, X.: PF-net: point fractal network for 3D point cloud completion. In: Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, pp. 7662–7670 (2020)
Liu, M., Sheng, L., Yang, S., Shao, J., Hu, S.M.: Morphing and sampling network for dense point cloud completion. In: Proceedings of the AAAI Conference on Artificial Intelligence, vol. 34, pp. 11596–11603 (2020)
Qi, C.R., Yi, L., Su, H., Guibas, L.J.: PointNet++: deep hierarchical feature learning on point sets in a metric space. arXiv preprint arXiv:1706.02413 (2017)
Tchapmi, L.P., Kosaraju, V., Rezatofighi, S.H., Reid, I., Savarese, S.: TopNet: structural point cloud decoder. In: Proceedings of the IEEE/CVF International Conference on Computer Vision (2019)
Vaswani, A., et al.: Attention is all you need. In: Advances in Neural Information Processing Systems, vol. 30 (2017)
Wang, X., M.H.A.J., Lee, G.H.: Cascaded refinement network for point cloud completion. In: Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, June 2020
Wang, X., Ang, M.H., Lee, G.H.: Voxel-based network for shape completion by leveraging edge generation. In: Proceedings of the IEEE/CVF International Conference on Computer Vision, pp. 13189–13198 (2021)
Wen, X., Han, Z., Cao, Y.P., Wan, P., Zheng, W., Liu, Y.S.: Cycle4Completion: unpaired point cloud completion using cycle transformation with missing region coding. In: Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, pp. 13075–13084 (2021)
Wen, X., Li, T., Han, Z., Liu, Y.S.: Point cloud completion by skip-attention network with hierarchical folding. In: Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, pp. 1939–1948 (2020)
Wen, X., et al.: PMP-net: point cloud completion by learning multi-step point moving paths. In: Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, pp. 7443–7452 (2021)
Xiang, P., et al.: SnowflakeNet: point cloud completion by snowflake point deconvolution with skip-transformer. In: Proceedings of the IEEE/CVF International Conference on Computer Vision, pp. 5499–5509 (October 2021)
Xie, H., Yao, H., Zhou, S., Mao, J., Zhang, S., Sun, W.: GRNet: gridding residual network for dense point cloud completion. In: Vedaldi, A., Bischof, H., Brox, T., Frahm, J.-M. (eds.) ECCV 2020. LNCS, vol. 12354, pp. 365–381. Springer, Cham (2020). https://doi.org/10.1007/978-3-030-58545-7_21
Yang, Y., Feng, C., Shen, Y., Tian, D.: FoldingNet: point cloud auto-encoder via deep grid deformation. In: Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, pp. 206–215 (2018)
Yu, X., Rao, Y., Wang, Z., Liu, Z., Lu, J., Zhou, J.: PoinTr: diverse point cloud completion with geometry-aware transformers. In: Proceedings of the IEEE/CVF International Conference on Computer Vision, pp. 12498–12507 (2021)
Yuan, W., Khot, T., Held, D., Mertz, C., Hebert, M.: PCN: point completion network. In: 3D Vision (3DV) (2018)
Zhang, W., Yan, Q., Xiao, C.: Detail preserved point cloud completion via separated feature aggregation. In: Vedaldi, A., Bischof, H., Brox, T., Frahm, J.-M. (eds.) ECCV 2020. LNCS, vol. 12370, pp. 512–528. Springer, Cham (2020). https://doi.org/10.1007/978-3-030-58595-2_31
Zhao, H., Jiang, L., Jia, J., Torr, P.H., Koltun, V.: Point transformer. In: Proceedings of the IEEE/CVF International Conference on Computer Vision, pp. 16259–16268 (2021)
Zhou, H., et al.: SeedFormer: patch seeds based point cloud completion with upsample transformer. In: Avidan, S., Brostow, G., Cissé, M., Farinella, G.M., Hassner, T. (eds.) ECCV 2022. LNCS, vol. 13663, pp. 416–432. Springer, Cham (2022). https://doi.org/10.1007/978-3-031-20062-5_24
Author information
Authors and Affiliations
Corresponding author
Editor information
Editors and Affiliations
Rights and permissions
Copyright information
© 2023 The Author(s), under exclusive license to Springer Nature Switzerland AG
About this paper
Cite this paper
Cheng, Z., Chen, X. (2023). Structure-Aware Point Cloud Completion. In: Lu, H., et al. Image and Graphics . ICIG 2023. Lecture Notes in Computer Science, vol 14357. Springer, Cham. https://doi.org/10.1007/978-3-031-46311-2_15
Download citation
DOI: https://doi.org/10.1007/978-3-031-46311-2_15
Published:
Publisher Name: Springer, Cham
Print ISBN: 978-3-031-46310-5
Online ISBN: 978-3-031-46311-2
eBook Packages: Computer ScienceComputer Science (R0)