《计算机应用》唯一官方网站 ›› 2026, Vol. 46 ›› Issue (8): 2594-2602.DOI: 10.11772/j.issn.1001-9081.2025070897

• 多媒体计算与计算机仿真 • 上一篇    下一篇

面向点云分类与分割的层次化旋转不变几何结构的表征学习方法

刘威1, 李维刚2(), 田志强2   

  1. 1.武汉科技大学 电子信息学院,武汉 430081
    2.武汉科技大学 人工智能与自动化学院,武汉 430081
  • 收稿日期:2025-08-07 修回日期:2025-10-15 接受日期:2025-10-15 发布日期:2025-11-05 出版日期:2026-08-10
  • 通讯作者: 李维刚
  • 作者简介:刘威(2001—),男,湖北孝感人,硕士研究生,主要研究方向:深度学习、点云数据处理
    李维刚(1977—),男,湖北咸宁人,教授,博士,主要研究方向:工业过程控制、人工智能、机器学习算法
    田志强(1996—),男,湖北武汉人,博士研究生,主要研究方向:计算机视觉。
  • 基金资助:
    湖北省科技人才服务企业项目(2024DJC077)

Representation learning method of hierarchical rotation-invariant geometric structure for point cloud classification and segmentation

Wei LIU1, Weigang LI2(), Zhiqiang TIAN2   

  1. 1.School of Electronic Information,Wuhan University of Science and Technology,Wuhan Hubei 430081,China
    2.School of Artificial Intelligence and Automation,Wuhan University of Science and Technology,Wuhan Hubei 430081,China
  • Received:2025-08-07 Revised:2025-10-15 Accepted:2025-10-15 Online:2025-11-05 Published:2026-08-10
  • Contact: Weigang LI
  • About author:LIU Wei, born in 2001, M. S. candidate. His research interests include deep learning, point cloud data processing.
    TIAN Zhiqiang, born in 1996, Ph. D. candidate. His research interests include computer vision.
  • Supported by:
    Hubei Provincial Science and Technology Talent Service Enterprise Project(2024DJC077)

摘要:

现有的点云深度学习方法能够有效处理固定视角下的点云数据;然而,在实际应用中,物体方向的变化会使点云描述受到旋转变换的影响,从而降低深度学习网络的识别精度。针对这一问题,提出一种面向点云分类与分割的层次化旋转不变几何结构的表征学习方法。首先,通过三角化的局部几何结构对点云样本进行建模,以在每个点的邻域内构建三角表面,提取描述欧氏空间与切平面几何关系的旋转不变特征;然后,将提取到的旋转不变特征通过卷积算子表达,并通过自注意力增强卷积聚合局部邻域结构,实现局部和全局信息的自适应融合,进一步提取精细的旋转不变特征并增强表达力和全局一致性;最后,引入层次化逆瓶颈残差模块(IRBlock),通过多级非线性映射和渐进式通道扩展,实现从浅层几何特征到深层语义特征的层次化特征融合,增强旋转不变特征的高阶表达能力和判断力,从而提升对复杂空间结构和多样旋转的表达和区分能力。实验结果表明,所提方法在ModelNet40数据集上实现了93.9%的整体准确率(OA),在ScanObjectNN数据集上实现了87.8%的OA,在ShapeNet数据集的分割任务中取得了82.3%的平均交并比(mIoU)。可见,所提方法具有良好的分类分割能力,同时兼具旋转不变性,表现出优异的鲁棒性和泛化能力。

关键词: 点云, 旋转不变性, 注意力机制, 分类分割, 深度学习

Abstract:

The existing point cloud deep learning methods can process point cloud data from fixed perspectives. However, in practical applications, changes in object directions affect the point cloud description by rotational transformations, thereby reducing recognition accuracy of the deep learning network. To address this issue, a representation learning method of hierarchical rotation-invariant geometric structure for point cloud classification and segmentation was proposed. First, point cloud samples were modeled using triangulated local geometric structure, so that a triangular surface was constructed within each neighborhood, and rotation-invariant features describing the geometric relationship between Euclidean space and the tangent plane were extracted. Then, the extracted rotation-invariant features were expressed through convolutional operators, and the self-attention enhanced convolutions were employed to aggregate local neighborhood structure, so as to achieve adaptive fusion of local and global information, thereby further extracting the refined rotation-invariant features as well as improving their expressiveness and global consistency. Finally, a hierarchical Inverted Residual Block (IRBlock) was introduced to fuse features from shallow geometric features to deep semantic features hierarchically through multi-level nonlinear mapping and progressive channel expansion, thereby enhancing the high-level expressiveness and discriminative ability of rotation-invariant features, and improving the ability to express and distinguish complex spatial structures and diverse rotations. Experimental results show that the proposed method achieves the Overall Accuracy (OA) of 93.9% on the ModelNet40 dataset, the OA of 87.8% on the ScanObjectNN dataset, and the mean Intersection over Union (mIoU) of 82.3% in the segmentation tasks of the ShapeNet dataset. It can be seen that the proposed method has strong classification and segmentation performance with rotation-invariance, and exhibits excellent robustness and generalization capability.

Key words: point cloud, rotation-invariance, attention mechanism, classification and segmentation, deep learning

中图分类号: