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. 62, Issue 10, 1000003(2025)

Review of Deep Learning-Based 3D Reconstruction

Wanyun Li, Yasheng Zhang, Yuqiang Fang*, Qinyu Zhu, and Xinli Zhu
Author Affiliations
  • Space Engineering University, Beijing 101416, China
    • show less
    AbstractGet PDF(in Chinese)
    Figures&Tables (14)
    Equations (3)
    References (95)
    Cited By (1)
    Tools
    Paper Information
    Topics
    Figures & Tables(14)
    Typical 3D shape representation methods. (a) Explicit representation; (b) implicit representation

    Fig. 1. Typical 3D shape representation methods. (a) Explicit representation; (b) implicit representation

    Download full size
    Schematic diagram of 3D ShapeNets[14]

    Fig. 2. Schematic diagram of 3D ShapeNets[14]

    Download full size
    3D-R2N2 network architecture[16]

    Fig. 3. 3D-R2N2 network architecture[16]

    Download full size
    Architecture of 3D Volume Transformer network[25]

    Fig. 4. Architecture of 3D Volume Transformer network[25]

    Download full size
    PointNet network architecture[27]

    Fig. 5. PointNet network architecture[27]

    Download full size
    Stereo2Point network architecture[33]

    Fig. 6. Stereo2Point network architecture[33]

    Download full size
    Pixel2Mesh network architecture[42]

    Fig. 7. Pixel2Mesh network architecture[42]

    Download full size
    VANet network architecture[46]

    Fig. 8. VANet network architecture[46]

    Download full size
    MISE flowchart

    Fig. 9. MISE flowchart

    Download full size
    Schematic of the continuous implicit function DeepSDF[12]

    Fig. 10. Schematic of the continuous implicit function DeepSDF[12]

    Download full size
    NeRF new perspective synthesis[13]

    Fig. 11. NeRF new perspective synthesis[13]

    Download full size
    Geo-NeuS network architecture[83]

    Fig. 12. Geo-NeuS network architecture[83]

    Download full size

    • Table 1. Summarization of explicit representation-based 3D reconstruction networks

      View table

      Table 1. Summarization of explicit representation-based 3D reconstruction networks

      MethodYearTraining processOutput 3D representationDataset
      Supervised or unsupervisedTraining network
      3D ShapeNets［14］2015SupervisedDBNVoxelModelNet
      TL-embedding［15］2016SupervisedE-DVoxelIKEA
      3D-R2N2［16］2016SupervisedE-DVoxelModelNet
      Pix2Vox［18］2019SupervisedE-DVoxelShapeNet， Pix3D
      Pix2Vox++［19］2020SupervisedE-DVoxelShapeNet， Pix3D
      P2VNet［20］2023SupervisedE-DVoxelThings3D
      OGN［22］2017SupervisedE-DVoxelShapeNet
      Adaptive O-CNN［23］2018SupervisedE-DVoxelShapeNet
      3D-RETR［24］2021SupervisedTransformerVoxelShapeNet
      3D Volume Transformer［25］2021SupervisedTransformerVoxelShapeNet， Pix3D
      PointNet［27］2017SupervisedE-DPointModelNet
      PointNet++［28］2017SupervisedE-DPointModelNet
      3D-LMNet［30］2018SupervisedE-DPointShapeNet
      Stereo2Point［33］2021UnsupervisedE-DPointStereoShapeNet
      Pixel2Mesh［42］2018SupervisedGCNMeshModelNet
      Pixel2Mesh++［43］2019SupervisedGCNMeshShapeNet
      Multi-view Pixel2Mesh++［45］2023SupervisedGCNMeshShapeNet
      VANet［46］2021SupervisedTransformerMeshShapeNet
      T-Pixel2Mesh［48］2024SupervisedTransformerMeshShapeNet

    • Table 2. Common 3D datasets

      View table

      Table 2. Common 3D datasets

      DatasetTypeNumber of typesNumber of 2D images3D model
      Quantity3D representation
      ShapeNetIndoor objects5551300CAD
      27012000CAD
      ModelNetIndoor objects660151128CAD
      Pix3DIndoor objects910069395CAD
      IKEAIndoor objects7800225CAD
      PASCAL 3D+Outdoor scene12308991015CAD
      ObjectNet3DMulti-scene1009012744147CAD
      DTUMulti-scene80254104235Point
      KITTIOutdoor scene314999Point
    Tools

    Get Citation

    Copy Citation Text

    Wanyun Li, Yasheng Zhang, Yuqiang Fang, Qinyu Zhu, Xinli Zhu. Review of Deep Learning-Based 3D Reconstruction[J]. Laser & Optoelectronics Progress, 2025, 62(10): 1000003

    Download Citation

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

    Category: Reviews

    Received: Sep. 24, 2024

    Accepted: Nov. 15, 2024

    Published Online: Apr. 23, 2025

    The Author Email: Yuqiang Fang (fangyuqiang@nudt.edu.cn)

    DOI:10.3788/LOP242030

    CSTR:32186.14.LOP242030

    Topics

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