融合可变形核和自注意力的点云分类分割边卷积网络

doi:10.19678/j.issn.1000-3428.0069284

摘要/Abstract

摘要：

点云数据具有无序性和离散分布的特点, 传统动态图卷积方法在处理点云数据时仍充满挑战, 无法准确地表示三维点间的特征对应关系。为此, 提出一种融合可变形核和自注意力的点云分类分割边卷积网络DKSA-Net, 该网络由DKConv (Deformable Kernels Edge Convolution)模块和SAConv (Self-Attention Edge Convolution)模块组成。通过融合可变形核与边卷积构建DKConv模块, 能够动态学习点的特征, 生成可变形核, 不会忽略不同特征之间的对应关系, 从而更好地处理不同特征之间的对应关系。引入自注意力机制, 并与边卷积结合构建SAConv模块, 能够对特征进行更细粒度的特征提取, 充分捕捉点云的重要特征, 增强模型的判别能力。实验结果表明, DKSA-Net在ModelNet40和ShapeNet数据集上取得出色性能, 分别达到93.4%的总体精度(OA)、90.7%的平均精度(mAcc)和86.1%的平均并交比(mIoU), 且有着较低模型复杂度和较好鲁棒性, 具有优秀的点云数据处理能力。

关键词: 可变形核, 自注意力, 点云分类, 点云分割, 深度学习

Abstract:

Owing to the unordered and discrete nature of point cloud data, traditional dynamic graph convolution method faces significant challenges in processing this data, making it difficult to accurately represent feature correspondences between 3D points. To address this issue, a network called DKSA-Net is proposed, which incorporates deformable kernels and self-attention. This network consists of two main modules: Deformable Kernels edge Convolution (DKConv) and Self-Attention edge Convolution (SAConv). By integrating deformable kernels with edge convolution to construct the DKConv module, the network can dynamically learn point features, generate deformable kernels, and maintain feature correspondences, thereby improving the handling of feature correspondences. In addition, by introducing the self-attention mechanism and combining it with edge convolution to construct the SAConv module, the network can perform finer-grained feature extraction, fully capture important point cloud features, and enhance the discriminative ability of the model. The experimental results show that DKSA-Net achieves excellent performance on the ModelNet40 and ShapeNet datasets, with an Overall Accuracy (OA) of 93.4%, an average Accuracy (mAcc) of 90.7%, and an average Intersection-over-Union (mIoU) of 86.1%. Furthermore, it demonstrates relatively low model complexity and high robustness, showcasing exceptional capabilities in processing point cloud data.

Key words: deformable kernel, self-attention, point cloud classification, point cloud segmentation, deep learning

陈思帆, 杨家志, 黄琳, 吕志玮, 沈露. 融合可变形核和自注意力的点云分类分割边卷积网络[J]. 计算机工程, 2025, 51(6): 146-154.

CHEN Sifan, YANG Jiazhi, HUANG Lin, Lü Zhiwei, SHEN Lu. Edge Convolutional Network for Point Cloud Classification and Segmentation Incorporated Deformable Kernel and Self-Attention[J]. Computer Engineering, 2025, 51(6): 146-154.

https://www.ecice06.com/CN/Y2025/V51/I6/146

图/表 16

图1 DKSA-Net网络模型架构

Fig.1 DKSA-Net network model structure

图2 DKConv模块

Fig.2 DKConv module

图3 可变形核

Fig.3 Deformable Kernels

图4 SAConv模块

Fig.4 SAConv module

图5 EdgeConv模块

Fig.5 EdgeConv module

图6 自注意力模块

Fig.6 Self-Attention module

图7 部件分割可视化结果

Fig.7 Component segmentation visualization results

图8 不同点云数量下的ModelNet40数据集分类精度

Fig.8 Classification accuracy on ModelNet40 dataset with different number of point clouds

图9 不同k值对网络模型的影响

Fig.9 Effect of different k values on the network model

参考文献 26

1	LI Y , IBANEZ-GUZMAN J . Lidar for autonomous driving: the principles, challenges, and trends for automotive lidar and perception systems. IEEE Signal Processing Magazine, 2020, 37 (4): 50- 61. doi: 10.1109/MSP.2020.2973615
2	AHMED S M, TAN Y Z, CHEW C M, et al. Edge and corner detection for unorganized 3D point clouds with application to robotic welding[C]//Proceedings of the 2018 IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS). New York, USA: ACM Press, 2018: 7350-7355.
3	RAHMAN M M , TAN Y H , XUE J , et al. Notice of violation of IEEE publication principles: recent advances in 3D object detection in the era of deep neural networks: a survey. IEEE Transactions on Image Processing, 2020, 29, 2947- 2962. doi: 10.1109/TIP.2019.2955239
4	张瑞菊, 周欣, 赵江洪, 等. 一种古建筑点云数据的语义分割算法. 武汉大学学报(信息科学版), 2020, 45 (5): 753- 759.
	ZHANG R J , ZHOU X , ZHAO J H , et al. A semantic segmentation algorithm of ancient building's point cloud data. Geomatics and Information Science of Wuhan University, 2020, 45 (5): 753- 759.
5	MATURANA D, SCHERER S, MATURANA D, et al. VoxNet: a 3D convolutional neural network for real-time object recognition[C]//Proceedings of the 2015 IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS). New York, USA: ACM Press, 2015: 922-928.
6	SU H, MAJI S, KALOGERAKIS E, et al. Multi-view convolutional neural networks for 3D shape recognition[C]//Proceedings of the IEEE International Conference on Computer Vision (ICCV). Washington D. C., USA: IEEE Press, 2015: 945-953.
7	QI C R, SU H, MO K, et al. PointNet: deep learning on point sets for 3D classification and segmentation[C]//Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition (CVPR). Washington D. C., USA: IEEE Press, 2017: 652-660.
8	QI C R, YI L, SU H, et al. PointNet++: deep hierarchical feature learning on point sets in a metric space[EB/OL]. [2023-10-12]. http://arxiv.org/abs/1706.02413.
9	YAN X, ZHENG C D, LI Z, et al. PointASNL: robust point clouds processing using nonlocal neural networks with adaptive sampling[C]//Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR). Washington D. C., USA: IEEE Press, 2020: 1-10.
10	LI J X, CHEN B M, LEE G H. SO-Net: self-organizing network for point cloud analysis[C]//Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition. Washington D. C., USA: IEEE Press, 2018: 9397-9406.
11	WANG Y , SUN Y B , LIU Z W , et al. Dynamic graph CNN for learning on point clouds. ACM Transactions on Graphics, 2019, 38 (5): 1- 12.
12	LI Y, BU R, SUN M, et al. PointCNN: convolution on X-transformed points[EB/OL]. [2023-10-12]. http://arxiv.org/abs/1801.07791.
13	LIN Z H, HUANG S Y, WANG Y F. Convolution in the cloud: learning deformable kernels in 3D graph convolution networks for point cloud analysis[C]//Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR). Washington D. C., USA: IEEE Press, 2020: 1800-1809.
14	VASWANI A, SHAZEER N, PARMAR N. Attention is all you need[EB/OL]. [2023-10-12]. https://arxiv.org/abs/1706.03762?context=cs.LG.
15	任欢, 王旭光. 注意力机制综述. 计算机应用, 2021, 41 (S1): 1- 6.
	REN H , WANG X G . Review of attention mechanism. Journal of Computer Applications, 2021, 41 (S1): 1- 6.
16	LOSHCHILOV I, HUTTER F. SGDR: stochastic gradient descent with warm restarts[EB/OL]. [2023-10-12]. https://arxiv.org/abs/1608.03983v5.
17	WU Z R, SONG S R, KHOSLA A, et al. 3D ShapeNets: a deep representation for volumetric shapes[C]//Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition (CVPR). Washington D. C., USA: IEEE Press, 2015: 1912-1920.
18	XU Y F, FAN T Q, XU M Y, et al. SpiderCNN: deep learning on point sets with parameterized convolutional filters[C]//Proceedings of European Conference on Computer Vision (ECCV). New York, USA: ACM Press, 2018: 90-105.
19	ATZMON M, MARON H, LIPMAN Y. Point convolutional neural networks by extension operators[EB/OL]. [2023-10-12]. https://arxiv.org/abs/1803.10091v1.
20	CHENG S L , CHEN X W , HE X W , et al. PRA-Net: point relation-aware network for 3D point cloud analysis. IEEE Transactions on Image Processing, 2021, 30, 4436- 4448. doi: 10.1109/TIP.2021.3072214
21	GAO Y B , LIU X B , LI J , et al. LFT-net: local feature transformer network for point clouds analysis. IEEE Transactions on Intelligent Transportation Systems, 2022, 24 (2): 1- 11.
22	田钰杰, 管有庆, 龚锐. 一种鲁棒的多特征点云分类分割深度神经网络. 计算机工程, 2021, 47 (11): 234- 240. doi: 10.19678/j.issn.1000-3428.0060004
	TIAN Y J , GUAN Y Q , GONG R . A robust deep neural network for multi-feature point cloud classification and segmentation. Computer Engineering, 2021, 47 (11): 234- 240. doi: 10.19678/j.issn.1000-3428.0060004
23	WIJAYA K T , PAEK D H , KONG S H . Advanced feature learning on point clouds using multi-resolution features and learnable pooling. Remote Sensing, 2024, 16 (11): 1835. doi: 10.3390/rs16111835
24	XIE Z Y , CHEN J Z , PENG B . Point clouds learning with attention-based graph convolution networks. Neurocomputing, 2020, 402, 245- 255. doi: 10.1016/j.neucom.2020.03.086
25	HAN X F, JIN, Y F, CHENG H X, et al. Dual transformer for point cloud analysis[EB/OL]. [2023-10-12]. https://arxiv.org/abs/2104.13044.
26	YI L , KIM V G , CEYLAN D , et al. A scalable active framework for region annotation in 3D shape collections. ACM Transactions on Graphics, 2016, 35 (6): 1- 12.

[1]	秦永旺, 张洋, 胡星, 刘胜, 李少青. 基于图注意力网络的门级网表功能识别[J]. 计算机工程, 2025, 51(6): 29-37.
[2]	庞鑫, 葛凤培, 李艳玲. 声景识音：数字化时代声学场景分类的探索与前沿[J]. 计算机工程, 2025, 51(6): 1-19.
[3]	廖丁丁, 刘俊峰, 曾君, 邱晓欢. 一种基于块平均正交权重修正的连续学习算法[J]. 计算机工程, 2025, 51(6): 57-64.
[4]	王培吉, 邹承明. 基于向量转换的卷积计算优化方法[J]. 计算机工程, 2025, 51(6): 74-82.
[5]	曹蓓, 赵奎. 基于双重情感和多特征融合的虚假新闻检测[J]. 计算机工程, 2025, 51(6): 193-203.
[6]	郝志峰, 黎阳霖, 许柏炎, 蔡瑞初. 面向跨域自然语言生成SQL语句的超图神经网络[J]. 计算机工程, 2025, 51(5): 114-123.
[7]	魏铭康, 李嘉楠, 韩林, 高伟, 赵荣彩, 王洪生. 面向深度学习编译器的多粒度量化框架支持与优化[J]. 计算机工程, 2025, 51(5): 62-72.
[8]	赵瑶谦, 滕奇志, 何小海, 税爱, 陈洪刚. 基于自注意力特征蒸馏的轻量级图像超分辨率重建[J]. 计算机工程, 2025, 51(5): 257-265.
[9]	庄紫薇, 朱俊国. 面向多源文本的越南语文本检错方法[J]. 计算机工程, 2025, 51(5): 93-102.
[10]	李丹丹, 李智, 郑龙, 张丽. 面向弥散张量图像的鲁棒可逆水印算法[J]. 计算机工程, 2025, 51(5): 279-287.
[11]	蒋杰平, 王明文. 基于时空置换注意力机制的残差行为识别模型[J]. 计算机工程, 2025, 51(4): 119-128.
[12]	杜晨阳, 张雪英, 黄丽霞, 李娟. 基于改进高效通道注意力机制的多特征语音情感识别[J]. 计算机工程, 2025, 51(4): 97-106.
[13]	刘云翔, 梁智超. 一种高效的连续时序图注意力网络的交通预测模型[J]. 计算机工程, 2025, 51(4): 350-359.
[14]	戴康佳, 徐慧英, 朱信忠, 李悉钰, 黄晓, 陈国强, 张志雄. YGL-SLAM: 动态场景下基于点和线的语义SLAM系统[J]. 计算机工程, 2025, 51(3): 95-104.
[15]	孙亭, 杨洁, 李家璇, 王耀宗. 面向弱光交通场景的YOLOv7道路标志检测算法优化[J]. 计算机工程, 2025, 51(3): 342-351.

选择文件类型/文献管理软件名称

选择包含的内容