《计算机应用》唯一官方网站 ›› 2024, Vol. 44 ›› Issue (4): 1086-1092.DOI: 10.11772/j.issn.1001-9081.2023050588
所属专题: 人工智能
收稿日期:
2023-05-16
修回日期:
2023-06-12
接受日期:
2023-06-21
发布日期:
2023-08-01
出版日期:
2024-04-10
通讯作者:
韩李涛
作者简介:
张鹏飞(1998—),男,河南驻马店人,硕士研究生,主要研究方向:三维点云数据处理、点云语义分割基金资助:
Pengfei ZHANG1, Litao HAN1,2(), Hengjian FENG1, Hongmei LI1
Received:
2023-05-16
Revised:
2023-06-12
Accepted:
2023-06-21
Online:
2023-08-01
Published:
2024-04-10
Contact:
Litao HAN
About author:
ZHANG Pengfei, born in 1998, M. S. candidate. His research interests include 3D point cloud data processing, point cloud semantic segmentation.Supported by:
摘要:
在基于深度学习的三维点云语义分割算法中,为了加强提取局部特征细粒度能力和学习不同局部邻域之间的长程依赖性,提出一种基于注意力机制和全局特征优化的神经网络。首先,通过加性注意力的形式设计单通道注意力(SCA)模块和点注意力(PA)模块,前者通过自适应调节单通道中各点特征加强对局部特征的分辨能力,后者通过调节单点特征向量之间的重要程度抑制无用特征并减少特征冗余;其次,加入全局特征聚合(GFA)模块,聚合各局部邻域特征,以捕获全局上下文信息,从而提高语义分割精度。实验结果表明,在点云数据集S3DIS上,所提网络的平均交并比(mIoU)相较于RandLA-Net(Random sampling and an effective Local feature Aggregator Network)提升了1.8个百分点,分割性能良好,具有较好的适应性。
中图分类号:
张鹏飞, 韩李涛, 冯恒健, 李洪梅. 基于注意力机制和全局特征优化的点云语义分割[J]. 计算机应用, 2024, 44(4): 1086-1092.
Pengfei ZHANG, Litao HAN, Hengjian FENG, Hongmei LI. Point cloud semantic segmentation based on attention mechanism and global feature optimization[J]. Journal of Computer Applications, 2024, 44(4): 1086-1092.
网络模型 | OA | mAcc | mIoU |
---|---|---|---|
PointNet[ | 79.3 | 49.0 | 41.1 |
DGCNN[ | 84.3 | 45.8 | 47.6 |
SPG[ | 86.4 | 66.5 | 58.0 |
HPEIN[ | 87.2 | 68.3 | 61.9 |
RandLA-Net[ | 87.2 | 71.5 | 62.4 |
BAAFNet[ | 87.0 | 71.2 | 62.3 |
MFNet[ | 88.5 | — | 62.9 |
MFF-Net[ | 87.1 | 72.4 | 63.0 |
本文网络 | 87.6 | 72.3 | 64.2 |
表1 不同网络在S3DIS数据集上的分割精度对比 (%)
Tab. 1 Comparison of segmentation accuracy among different networks on S3DIS dataset
网络模型 | OA | mAcc | mIoU |
---|---|---|---|
PointNet[ | 79.3 | 49.0 | 41.1 |
DGCNN[ | 84.3 | 45.8 | 47.6 |
SPG[ | 86.4 | 66.5 | 58.0 |
HPEIN[ | 87.2 | 68.3 | 61.9 |
RandLA-Net[ | 87.2 | 71.5 | 62.4 |
BAAFNet[ | 87.0 | 71.2 | 62.3 |
MFNet[ | 88.5 | — | 62.9 |
MFF-Net[ | 87.1 | 72.4 | 63.0 |
本文网络 | 87.6 | 72.3 | 64.2 |
模型 | 交并比 | ||||||||||||
---|---|---|---|---|---|---|---|---|---|---|---|---|---|
天花板 | 地板 | 墙壁 | 梁 | 柱子 | 窗户 | 门 | 桌子 | 椅子 | 沙发 | 书柜 | 板子 | 杂物 | |
PointNet | 88.8 | 97.3 | 69.8 | 0.1 | 3.9 | 46.3 | 10.8 | 59.0 | 52.6 | 5.9 | 40.3 | 26.4 | 33.2 |
DGCNN | 92.7 | 97.5 | 74.8 | 0.0 | 11.7 | 50.7 | 23.7 | 66.3 | 69.5 | 8.5 | 48.7 | 31.5 | 42.6 |
SPG | 89.4 | 96.9 | 78.1 | 0.0 | 42.8 | 48.9 | 61.6 | 84.7 | 75.4 | 69.8 | 52.6 | 2.1 | 52.2 |
HPEIN | 91.5 | 98.2 | 81.4 | 0.0 | 23.3 | 65.3 | 40.0 | 75.5 | 87.7 | 58.5 | 67.8 | 65.6 | 49.4 |
RandLA-Net | 91.1 | 95.6 | 80.2 | 0.0 | 24.7 | 62.3 | 47.7 | 76.2 | 83.7 | 60.2 | 71.1 | 65.7 | 53.8 |
MFNet | 91.2 | 98.2 | 80.4 | 0.0 | 35.1 | 57.2 | 46.6 | 77.4 | 86.0 | 65.6 | 65.4 | 66.2 | 47.6 |
BAAFNet | 92.0 | 97.9 | 80.9 | 0.0 | 30.9 | 58.2 | 47.6 | 78.1 | 87.2 | 52.4 | 67.0 | 67.3 | 51.0 |
MFF-Net | 93.0 | 98.0 | 81.4 | 0.0 | 26.4 | 59.9 | 46.2 | 78.9 | 87.1 | 65.8 | 65.8 | 64.9 | 51.0 |
本文网络 | 91.8 | 97.4 | 80.9 | 0.0 | 23.6 | 61.1 | 49.5 | 78.6 | 86.6 | 73.1 | 71.1 | 68.6 | 52.8 |
表2 S3DIS数据集上各类别实验结果对比 (%)
Tab. 2 Comparison of experimental results for various categories on S3DIS dataset
模型 | 交并比 | ||||||||||||
---|---|---|---|---|---|---|---|---|---|---|---|---|---|
天花板 | 地板 | 墙壁 | 梁 | 柱子 | 窗户 | 门 | 桌子 | 椅子 | 沙发 | 书柜 | 板子 | 杂物 | |
PointNet | 88.8 | 97.3 | 69.8 | 0.1 | 3.9 | 46.3 | 10.8 | 59.0 | 52.6 | 5.9 | 40.3 | 26.4 | 33.2 |
DGCNN | 92.7 | 97.5 | 74.8 | 0.0 | 11.7 | 50.7 | 23.7 | 66.3 | 69.5 | 8.5 | 48.7 | 31.5 | 42.6 |
SPG | 89.4 | 96.9 | 78.1 | 0.0 | 42.8 | 48.9 | 61.6 | 84.7 | 75.4 | 69.8 | 52.6 | 2.1 | 52.2 |
HPEIN | 91.5 | 98.2 | 81.4 | 0.0 | 23.3 | 65.3 | 40.0 | 75.5 | 87.7 | 58.5 | 67.8 | 65.6 | 49.4 |
RandLA-Net | 91.1 | 95.6 | 80.2 | 0.0 | 24.7 | 62.3 | 47.7 | 76.2 | 83.7 | 60.2 | 71.1 | 65.7 | 53.8 |
MFNet | 91.2 | 98.2 | 80.4 | 0.0 | 35.1 | 57.2 | 46.6 | 77.4 | 86.0 | 65.6 | 65.4 | 66.2 | 47.6 |
BAAFNet | 92.0 | 97.9 | 80.9 | 0.0 | 30.9 | 58.2 | 47.6 | 78.1 | 87.2 | 52.4 | 67.0 | 67.3 | 51.0 |
MFF-Net | 93.0 | 98.0 | 81.4 | 0.0 | 26.4 | 59.9 | 46.2 | 78.9 | 87.1 | 65.8 | 65.8 | 64.9 | 51.0 |
本文网络 | 91.8 | 97.4 | 80.9 | 0.0 | 23.6 | 61.1 | 49.5 | 78.6 | 86.6 | 73.1 | 71.1 | 68.6 | 52.8 |
网络模型 | OA | mIoU |
---|---|---|
RandLA-Net[ | 90.0 | 70.6 |
本文网络 | 90.0 | 73.5 |
表3 不同网络在Semantic3D数据集上的分割精度对比 (%)
Tab. 3 Comparison of segmentation accuracy among different networks on Semantic3D dataset
网络模型 | OA | mIoU |
---|---|---|
RandLA-Net[ | 90.0 | 70.6 |
本文网络 | 90.0 | 73.5 |
模型 | 交并比 | |||||||
---|---|---|---|---|---|---|---|---|
人造 地形 | 自然 地形 | 高植被 | 低植被 | 建筑物 | 人造 景观 | 扫描 伪影 | 汽车 | |
RandLA-Net | 95.4 | 85.0 | 84.7 | 33.1 | 84.4 | 26.9 | 63.4 | 92.3 |
本文网络 | 93.5 | 96.4 | 89.3 | 53.2 | 86.6 | 30.4 | 79.9 | 79.0 |
表4 Semantic3D数据集上各类别实验结果对比 (%)
Tab. 4 Comparison of experimental results for various categories on Semantic3D dataset
模型 | 交并比 | |||||||
---|---|---|---|---|---|---|---|---|
人造 地形 | 自然 地形 | 高植被 | 低植被 | 建筑物 | 人造 景观 | 扫描 伪影 | 汽车 | |
RandLA-Net | 95.4 | 85.0 | 84.7 | 33.1 | 84.4 | 26.9 | 63.4 | 92.3 |
本文网络 | 93.5 | 96.4 | 89.3 | 53.2 | 86.6 | 30.4 | 79.9 | 79.0 |
网络 | PA | SCA | SDPA | GFA | mIoU |
---|---|---|---|---|---|
1 | 62.4 | ||||
2 | √ | 63.1 | |||
3 | √ | 63.0 | |||
4 | √ | √ | 64.0 | ||
5 | √ | 61.5 | |||
6 | √ | √ | √ | 64.2 |
表5 消融实验结果 (%)
Tab. 5 Results of ablation experiments
网络 | PA | SCA | SDPA | GFA | mIoU |
---|---|---|---|---|---|
1 | 62.4 | ||||
2 | √ | 63.1 | |||
3 | √ | 63.0 | |||
4 | √ | √ | 64.0 | ||
5 | √ | 61.5 | |||
6 | √ | √ | √ | 64.2 |
网络模型 | 计算量/MB | 总时间/s |
---|---|---|
RandLA-Net | 26.3 | 4.06 |
RandLA-Net+PA | 28.4 | 4.20 |
RandLA-Net+PA+SCA | 32.7 | 4.76 |
RandLA-Net+PA+SCA+GFA | 44.2 | 5.40 |
表6 添加不同模块的计算量和总时间对比结果
Tab. 6 Computational amount and total time comparison results of adding different modules
网络模型 | 计算量/MB | 总时间/s |
---|---|---|
RandLA-Net | 26.3 | 4.06 |
RandLA-Net+PA | 28.4 | 4.20 |
RandLA-Net+PA+SCA | 32.7 | 4.76 |
RandLA-Net+PA+SCA+GFA | 44.2 | 5.40 |
1 | LIN Y, VOSSELMAN G, CAO Y, et al. Local and global encoder network for semantic segmentation of airborne laser scanning point clouds[J]. ISPRS Journal of Photogrammetry and Remote Sensing, 2021, 176: 151-168. 10.1016/j.isprsjprs.2021.04.016 |
2 | ZHANG J, LIN X, NING X. SVM-based classification of segmented airborne LiDAR point clouds in urban areas[J]. Remote Sensing, 2013, 5(8): 3749-3775. 10.3390/rs5083749 |
3 | CHEHATA N, GUO L, MALLET C. Airborne LIDAR feature selection for urban classification using random forests[C]// Laser Scanning 2009. Paris: ISPRS.org, 2009: 207-212. 10.1109/icip.2009.5414234 |
4 | ZHUANG Y, LIU Y, HE G, et al. Contextual classification of 3D laser points with conditional random fields in urban environments[C]// Proceedings of the 2015 IEEE/RSJ International Conference on Intelligent Robots and Systems. Piscataway: IEEE, 2015: 3908-3913. 10.1109/iros.2015.7353927 |
5 | LU Y, RASMUSSEN C. Simplified Markov random fields for efficient semantic labeling of 3D point clouds[C]// Proceedings of the 2012 IEEE/RSJ International Conference on Intelligent Robots and Systems. Piscataway: IEEE, 2012: 2690-2697. 10.1109/iros.2012.6386039 |
6 | TONG G, PENG H, SHAO Y, et al. ASCNet: 3D object detection from point cloud based on adaptive spatial context features[J]. Neurocomputing, 2022, 475: 89-101. 10.1016/j.neucom.2021.12.061 |
7 | LIU M, ZHOU Q, ZHAO H, et al. Prototype-voxel contrastive learning for LiDAR point cloud panoptic segmentation[C]// Proceedings of the 2022 International Conference on Robotics and Automation. Piscataway: IEEE, 2022: 9243-9250. 10.1109/icra46639.2022.9811638 |
8 | 郝雯,王红霄,汪洋.结合空间注意力与形状特征的三维点云语义分割[J].激光与光电子学进展,2022,59(8):0828004. 10.3788/LOP202259.0828004 |
HAO W, WANG H X, WANG Y. Semantic segmentation of three-dimensional point cloud based on spatial attention and shape feature[J]. Laser & Optoelectronics Progress, 2022, 59(8):0828004. 10.3788/LOP202259.0828004 | |
9 | SU H, MAJI S, KALOGERAKIS E, et al. Multi-view convolutional neural networks for 3D shape recognition[C]// Proceedings of the 2015 IEEE International Conference on Computer Vision. Piscataway: IEEE, 2015: 945-953. 10.1109/iccv.2015.114 |
10 | MILIOTO A, VIZZO I, BEHLEY J, et al. RangeNet++: fast and accurate LiDAR semantic segmentation[C]// Proceedings of the 2019 IEEE/RSJ International Conference on Intelligent Robots and Systems. Piscataway: IEEE, 2019: 4213-4220. 10.1109/iros40897.2019.8967762 |
11 | GUO Y, WANG H, HU Q, et al. Deep learning for 3D point clouds: a survey[J]. IEEE Transactions on Pattern Analysis and Machine Intelligence, 2020, 43(12): 4338-4364. 10.1109/tpami.2020.3005434 |
12 | HUANG J, YOU S. Point cloud labeling using 3D convolutional neural network[C]// Proceedings of the 2016 23rd International Conference on Pattern Recognition. Piscataway: IEEE, 2016: 2670-2675. 10.1109/icpr.2016.7900038 |
13 | RIEGLER G, OSMAN ULUSOY A, GEIGER A. OctNet: learning deep 3D representations at high resolutions[C]// Proceedings of the 2017 IEEE Conference on Computer Vision and Pattern Recognition. Piscataway: IEEE Computer Society, 2017: 6620-6629. 10.1109/cvpr.2017.701 |
14 | 文沛, 程英蕾, 余旺盛. 基于深度学习的点云分类方法综述[J]. 激光与光电子学进展, 2021, 58(16):1600003. 10.3788/lop202158.1600003 |
WEN P, CHENG Y L, YU W S. Point cloud classification methods based on deep learning: a review[J]. Laser & Optoelectronics Progress, 2021,58(16):1600003. 10.3788/lop202158.1600003 | |
15 | TCHAPMI L, CHOY C, ARMENI I, et al. SEGCloud: semantic segmentation of 3D point clouds[C]// Proceedings of the 2017 International Conference on 3D Vision. Piscataway: IEEE, 2017: 537-547. 10.1109/3dv.2017.00067 |
16 | QI C R, SU H, MO K, et al. PointNet: deep learning on point sets for 3D classification and segmentation[C]// Proceedings of the 2017 IEEE Conference on Computer Vision and Pattern Recognition. Piscataway: IEEE Computer Society, 2017: 652-660. 10.1109/cvpr.2017.16 |
17 | QI C R, YI L, SU H, et al. PointNet++: deep hierarchical feature learning on point sets in a metric space[C]// Proceedings of the 31st International Conference on Neural Information Processing Systems. Red Hook: Curran Associates Inc., 2017: 5105-5114. |
18 | HU Q, YANG B, XIE L, et al. RandLA-Net: efficient semantic segmentation of large-scale point clouds[C]// Proceedings of the 2020 IEEE/CVF Conference on Computer Vision and Pattern Recognition. Piscataway: IEEE Computer Society, 2020: 11108-11117. 10.1109/cvpr42600.2020.01112 |
19 | QIU S, ANWAR S, BARNES N. Semantic segmentation for real point cloud scenes via bilateral augmentation and adaptive fusion[C]// Proceedings of the 2021 IEEE/CVF Conference on Computer Vision and Pattern Recognition. Piscataway: IEEE Computer Society, 2021: 1757-1767. 10.1109/cvpr46437.2021.00180 |
20 | SHAO Y, TONG G, PENG H. Mining local geometric structure for large-scale 3D point clouds semantic segmentation[J]. Neurocomputing, 2022, 500: 191-202. 10.1016/j.neucom.2022.05.060 |
21 | LI Y, LIN Q, ZHANG Z, et al. MFNet: multi-level feature extraction and fusion network for large-scale point cloud classification[J]. Remote Sensing, 2022, 14(22): 5707. 10.3390/rs14225707 |
22 | WANG Y, SUN Y, LIU Z, et al. Dynamic graph CNN for learning on point clouds[J]. ACM Transactions on Graphics, 2019, 38(5): No. 146. 10.1145/3326362 |
23 | LANDRIEU L, SIMONOVSKY M. Large-scale point cloud semantic segmentation with superpoint graphs[C]// Proceedings of the 2018 IEEE/CVF Conference on Computer Vision and Pattern Recognition. Piscataway: IEEE Computer Society, 2018: 4558-4567. 10.1109/cvpr.2018.00479 |
24 | JIANG L, ZHAO H, LIU S, et al. Hierarchical point-edge interaction network for point cloud semantic segmentation[C]// Proceedings of the 2019 IEEE/CVF International Conference on Computer Vision. Piscataway: IEEE Computer Society, 2019: 10433-10441. 10.1109/iccv.2019.01053 |
25 | HUANG R, XU Y, STILLA U. GraNet: global relation-aware attentional network for semantic segmentation of ALS point clouds[J]. ISPRS Journal of Photogrammetry and Remote Sensing, 2021, 177: 1-20. 10.1016/j.isprsjprs.2021.04.017 |
26 | FAN S, DONG Q, ZHU F, et al. SCF-Net: learning spatial contextual features for large-scale point cloud segmentation[C]// Proceedings of the 2021 IEEE/CVF Conference on Computer Vision and Pattern Recognition. Piscataway: IEEE Computer Society, 2021: 14504-14513. 10.1109/cvpr46437.2021.01427 |
27 | ARMENI I, SENER O, ZAMIR A R, et al. 3D semantic parsing of large-scale indoor spaces[C]// Proceedings of the 2016 IEEE Conference on Computer Vision and Pattern Recognition. Piscataway: IEEE, 2016: 1534-1543. 10.1109/cvpr.2016.170 |
28 | HACKEL T, SAVINOV N, LADICKY L, et al. Semantic3D. net: a new large-scale point cloud classification benchmark[J]. ISPRS Annals of the Photogrammetry, Remote Sensing and Spatial Information Sciences, 2017, Ⅵ-1/W1: 91-98. 10.5194/isprs-annals-iv-1-w1-91-2017 |
29 | 郝雯,汪洋,魏海南.基于多特征融合的点云场景语义分割[J]. 计算机应用, 2023, 43(10): 3202-3208. |
HAO W, WANG Y, WEI H N. Semantic segmentation of point cloud scenes based on multi-feature fusion[J]. Journal of Computer Applications, 2023, 43(10): 3202-3208. |
[1] | 潘烨新, 杨哲. 基于多级特征双向融合的小目标检测优化模型[J]. 《计算机应用》唯一官方网站, 2024, 44(9): 2871-2877. |
[2] | 赵志强, 马培红, 黑新宏. 基于双重注意力机制的人群计数方法[J]. 《计算机应用》唯一官方网站, 2024, 44(9): 2886-2892. |
[3] | 李顺勇, 李师毅, 胥瑞, 赵兴旺. 基于自注意力融合的不完整多视图聚类算法[J]. 《计算机应用》唯一官方网站, 2024, 44(9): 2696-2703. |
[4] | 秦璟, 秦志光, 李发礼, 彭悦恒. 基于概率稀疏自注意力神经网络的重性抑郁疾患诊断[J]. 《计算机应用》唯一官方网站, 2024, 44(9): 2970-2974. |
[5] | 王熙源, 张战成, 徐少康, 张宝成, 罗晓清, 胡伏原. 面向手术导航3D/2D配准的无监督跨域迁移网络[J]. 《计算机应用》唯一官方网站, 2024, 44(9): 2911-2918. |
[6] | 李力铤, 华蓓, 贺若舟, 徐况. 基于解耦注意力机制的多变量时序预测模型[J]. 《计算机应用》唯一官方网站, 2024, 44(9): 2732-2738. |
[7] | 黄云川, 江永全, 黄骏涛, 杨燕. 基于元图同构网络的分子毒性预测[J]. 《计算机应用》唯一官方网站, 2024, 44(9): 2964-2969. |
[8] | 薛凯鹏, 徐涛, 廖春节. 融合自监督和多层交叉注意力的多模态情感分析网络[J]. 《计算机应用》唯一官方网站, 2024, 44(8): 2387-2392. |
[9] | 汪雨晴, 朱广丽, 段文杰, 李书羽, 周若彤. 基于交互注意力机制的心理咨询文本情感分类模型[J]. 《计算机应用》唯一官方网站, 2024, 44(8): 2393-2399. |
[10] | 高鹏淇, 黄鹤鸣, 樊永红. 融合坐标与多头注意力机制的交互语音情感识别[J]. 《计算机应用》唯一官方网站, 2024, 44(8): 2400-2406. |
[11] | 刘禹含, 吉根林, 张红苹. 基于骨架图与混合注意力的视频行人异常检测方法[J]. 《计算机应用》唯一官方网站, 2024, 44(8): 2551-2557. |
[12] | 李钟华, 白云起, 王雪津, 黄雷雷, 林初俊, 廖诗宇. 基于图像增强的低照度人脸检测[J]. 《计算机应用》唯一官方网站, 2024, 44(8): 2588-2594. |
[13] | 莫尚斌, 王文君, 董凌, 高盛祥, 余正涛. 基于多路信息聚合协同解码的单通道语音增强[J]. 《计算机应用》唯一官方网站, 2024, 44(8): 2611-2617. |
[14] | 顾焰杰, 张英俊, 刘晓倩, 周围, 孙威. 基于时空多图融合的交通流量预测[J]. 《计算机应用》唯一官方网站, 2024, 44(8): 2618-2625. |
[15] | 石乾宏, 杨燕, 江永全, 欧阳小草, 范武波, 陈强, 姜涛, 李媛. 面向空气质量预测的多粒度突变拟合网络[J]. 《计算机应用》唯一官方网站, 2024, 44(8): 2643-2650. |
阅读次数 | ||||||
全文 |
|
|||||
摘要 |
|
|||||