Semantic Segmentation of 3D Point Cloud Based on Spatial Eight-Quadrant Kernel Convolution

被引:4
|
作者
Liu, Liman [1 ]
Yu, Jinjin [1 ]
Tan, Longyu [1 ]
Su, Wanjuan [2 ]
Zhao, Lin [2 ]
Tao, Wenbing [2 ]
机构
[1] South Cent Univ Nationalities, Sch Biomed Engn, Wuhan 430074, Peoples R China
[2] Huazhong Univ Sci & Technol, Sch Artificial Intelligence & Automat, Natl Key Lab Sci & Technol Multispectral Informat, Wuhan 430074, Peoples R China
基金
中国国家自然科学基金;
关键词
spatial eight-quadrant kernel convolution; 3D point cloud; semantic segmentation; indoor scene;
D O I
10.3390/rs13163140
中图分类号
X [环境科学、安全科学];
学科分类号
08 ; 0830 ;
摘要
In order to deal with the problem that some existing semantic segmentation networks for 3D point clouds generally have poor performance on small objects, a Spatial Eight-Quadrant Kernel Convolution (SEQKC) algorithm is proposed to enhance the ability of the network for extracting fine-grained features from 3D point clouds. As a result, the semantic segmentation accuracy of small objects in indoor scenes can be improved. To be specific, in the spherical space of the point cloud neighborhoods, a kernel point with attached weights is constructed in each octant, the distances between the kernel point and the points in its neighborhood are calculated, and the distance and the kernel points' weights are used together to weight the point cloud features in the neighborhood space. In this case, the relationship between points are modeled, so that the local fine-grained features of the point clouds can be extracted by the SEQKC. Based on the SEQKC, we design a downsampling module for point clouds, and embed it into classical semantic segmentation networks (PointNet++, PointSIFT and PointConv) for semantic segmentation. Experimental results on benchmark dataset ScanNet V2 show that SEQKC-based PointNet++, PointSIFT and PointConv outperform the original networks about 1.35-2.12% in terms of MIoU, and they effectively improve the semantic segmentation performance of the networks for small objects of indoor scenes, e.g., the segmentation accuracy of small object "picture" is improved from 0.70% of PointNet++ to 10.37% of SEQKC-PointNet++.
引用
收藏
页数:16
相关论文
共 50 条
  • [41] 3D point cloud classification and segmentation based on dual attention and weighted dynamic graph convolution
    Xiao, Jian
    Wang, Xiaohong
    Li, Wei
    Yang, Yifei
    Luo, Ji
    Guangxue Jingmi Gongcheng/Optics and Precision Engineering, 2024, 32 (18): : 2823 - 2835
  • [42] Kernel Point Convolution LSTM Networks for Radar Point Cloud Segmentation
    Nobis, Felix
    Fent, Felix
    Betz, Johannes
    Lienkamp, Markus
    APPLIED SCIENCES-BASEL, 2021, 11 (06):
  • [43] Linking Points With Labels in 3D: A Review of Point Cloud Semantic Segmentation
    Xie, Yuxing
    Tian, Jiaojiao
    Zhu, Xiao Xiang
    IEEE GEOSCIENCE AND REMOTE SENSING MAGAZINE, 2020, 8 (04) : 38 - 59
  • [44] Crossmodal Few-shot 3D Point Cloud Semantic Segmentation
    Zhao, Ziyu
    Wu, Zhenyao
    Wu, Xinyi
    Zhang, Canyu
    Wang, Song
    PROCEEDINGS OF THE 30TH ACM INTERNATIONAL CONFERENCE ON MULTIMEDIA, MM 2022, 2022, : 4760 - 4768
  • [45] LONet: Local Optimization Network for 3D point cloud semantic segmentation
    Su, Shengbin
    Lu, Jian
    Chen, Xiaogai
    Zhang, Kaibing
    Zhou, Jian
    DIGITAL SIGNAL PROCESSING, 2024, 154
  • [46] AttAN: Attention Adversarial Networks for 3D Point Cloud Semantic Segmentation
    Zhang, Gege
    Ma, Qinghua
    Jiao, Licheng
    Liu, Fang
    Sun, Qigong
    PROCEEDINGS OF THE TWENTY-NINTH INTERNATIONAL JOINT CONFERENCE ON ARTIFICIAL INTELLIGENCE, 2020, : 789 - 796
  • [47] MHNet: Multiscale Hierarchical Network for 3D Point Cloud Semantic Segmentation
    Liang, Xiaoli
    Fu, Zhongliang
    IEEE ACCESS, 2019, 7 : 173999 - 174012
  • [48] Boosting Lidar 3D Object Detection with Point Cloud Semantic Segmentation
    Zhang, Xuchong
    Min, Chong
    Jia, Yijie
    Chen, Liming
    Zhang, Jingmin
    Sun, Hongbin
    2023 IEEE/RSJ INTERNATIONAL CONFERENCE ON INTELLIGENT ROBOTS AND SYSTEMS (IROS), 2023, : 7614 - 7621
  • [49] Improving Point Cloud Semantic Segmentation by Learning 3D Object Detection
    Unal, Ozan
    Van Gool, Luc
    Dai, Dengxin
    2021 IEEE WINTER CONFERENCE ON APPLICATIONS OF COMPUTER VISION WACV 2021, 2021, : 2949 - 2958
  • [50] A Depth Image Fusion Network for 3D Point Cloud Semantic Segmentation
    Wang, Zhou
    Jia, Zixi
    Lyu, Ao
    Wang, Yating
    Sun, Changsheng
    Liu, Yongxin
    2019 9TH IEEE ANNUAL INTERNATIONAL CONFERENCE ON CYBER TECHNOLOGY IN AUTOMATION, CONTROL, AND INTELLIGENT SYSTEMS (IEEE-CYBER 2019), 2019, : 849 - 853