Journals Articles Conferences News About CLP
Submit a paper Email Alert
Search
Search
Articles
Journals
News
Advanced Search
Top Searches
2D MaterialsTransformation opticsQuantum PhotonicsSmall lasersSemiconductor UV PhotonicsSupersymmetric microring laser arrays

Laser & Optoelectronics Progress, Volume. 61, Issue 10, 1015002(2024)

Attention-Based Multi-Stage Network for Point Cloud Completion

Xiyang Yin1, Pei Zhou1,2, and Jiangping Zhu1,2、*
Author Affiliations
  • 1College of Computer Science, Sichuan University, Chengdu 610065, Sichuan, China
  • 2National Key Laboratory of Fundamental Science on Synthetic Vision, Sichuan University, Chengdu 610065, Sichuan, China
    • show less
    AbstractGet PDF(in Chinese)
    Figures&Tables (12)
    Equations (9)
    References (25)
    Tools
    Paper Information
    Topics
    Figures & Tables(12)
    Overall architecture of AMCNet

    Fig. 1. Overall architecture of AMCNet

    Download full size
    Illustration of the aggregation of local features in pyramid feature extractor

    Fig. 2. Illustration of the aggregation of local features in pyramid feature extractor

    Download full size
    GAB. (a) Improved cross-attention module; (b) self-attention module; (c) channel attention SELayer module (⊕ denotes element-wise addition and ⊙ denotes Hadamard product)

    Fig. 3. GAB. (a) Improved cross-attention module; (b) self-attention module; (c) channel attention SELayer module (⊕ denotes element-wise addition and ⊙ denotes Hadamard product)

    Download full size
    Structure of the point generator

    Fig. 4. Structure of the point generator

    Download full size
    Visualization of completion results of different networks on PCN dataset

    Fig. 5. Visualization of completion results of different networks on PCN dataset

    Download full size
    Visualization of completion results of different networks in terms of the chair class

    Fig. 6. Visualization of completion results of different networks in terms of the chair class

    Download full size
    Visualization of completion results at different resolutions of input

    Fig. 7. Visualization of completion results at different resolutions of input

    Download full size
    Visualization of completion results of different point cloud block size

    Fig. 8. Visualization of completion results of different point cloud block size

    Download full size

    • Table 1. Point cloud completion comparison on PCN dataset in terms of CD(lower is better)

      View table

      Table 1. Point cloud completion comparison on PCN dataset in terms of CD(lower is better)

      ModelCD /103
      AveragePlaneCabinetCarChairLampCouchTableBoat
      FoldingNet14.319.4915.8012.6115.5516.4115.9713.6514.99
      PCN9.645.5022.7010.638.7011.0011.3411.688.59
      GRNet8.836.4510.379.459.417.9610.515.448.04
      PMP-Net8.735.6511.249.649.516.9510.838.727.25
      PoinTr8.384.7510.478.689.397.7510.937.757.29
      SnowflakeNet7.214.299.168.087.896.079.236.556.40
      PointAttN6.863.879.007.637.435.908.686.326.09
      Ours6.453.588.747.366.865.288.325.885.69

    • Table 2. Effect of the resolution of the input point cloud

      View table

      Table 2. Effect of the resolution of the input point cloud

      Resolution20481024512256
      CD6.456.546.667.23

    • Table 3. Effect of point cloud block size

      View table

      Table 3. Effect of point cloud block size

      SizeK=16K=32K=64
      CD6.486.456.47

    • Table 4. Effect of attention module

      View table

      Table 4. Effect of attention module

      ModelSELayerSkip connectionCD
      A6.61
      B√6.58
      C√6.49
      AMCNet√√6.45
    Tools

    Get Citation

    Copy Citation Text

    Xiyang Yin, Pei Zhou, Jiangping Zhu. Attention-Based Multi-Stage Network for Point Cloud Completion[J]. Laser & Optoelectronics Progress, 2024, 61(10): 1015002

    Download Citation

    EndNote(RIS)BibTexPlain Text
    Set citation alerts for article
    Save article for my favorites
    Paper Information

    Category: Machine Vision

    Received: Jul. 19, 2023

    Accepted: Oct. 9, 2023

    Published Online: Mar. 20, 2024

    The Author Email: Jiangping Zhu (zjp16@scu.edu.cn)

    DOI:10.3788/LOP231758

    CSTR:32186.14.LOP231758

    Topics

    laser devices and laser physics
    Lasers and Laser Optics
    Laser physics
    laser manufacturing
    Instrumentation, Measurement and Metrology