OmniSDF: Scene Reconstruction using Omnidirectional Signed Distance Functions and Adaptive Binoctrees

被引:0
|
作者
Kim, Hakyeong [1 ]
Meuleman, Andreas [2 ]
Jang, Hyeonjoong [1 ]
Tompkin, James [3 ]
Kim, Min H. [1 ]
机构
[1] Korea Adv Inst Sci & Technol, Daejeon, South Korea
[2] INRIA, Paris, France
[3] Brown Univ, Providence, RI USA
关键词
D O I
10.1109/CVPR52733.2024.01912
中图分类号
TP18 [人工智能理论];
学科分类号
081104 ; 0812 ; 0835 ; 1405 ;
摘要
We present a method to reconstruct indoor and outdoor static scene geometry and appearance from an omnidirectional video moving in a small circular sweep. This setting is challenging because of the small baseline and large depth ranges, making it difficult to find ray crossings. To better constrain the optimization, we estimate geometry as a signed distance field within a spherical binoctree data structure and use a complementary efficient tree traversal strategy based on a breadth-first search for sampling. Unlike regular grids or trees, the shape of this structure well-matches the camera setting, creating a better memory-quality trade-off. From an initial depth estimate, the binoctree is adaptively subdivided throughout the optimization; previous methods use a fixed depth that leaves the scene undersampled. In comparison with three neural optimization methods and two non-neural methods, ours shows decreased geometry error on average, especially in a detailed scene, while significantly reducing the required number of voxels to represent such details.
引用
收藏
页码:20227 / 20236
页数:10
相关论文
共 44 条
  • [1] Real-time Reconstruction of Depth Sequences Using Signed Distance Functions
    Morgan, John P., Jr.
    Tutwiler, Richard L.
    SIGNAL PROCESSING, SENSOR/INFORMATION FUSION, AND TARGET RECOGNITION XXIII, 2014, 9091
  • [2] Multi-View Reconstruction using Signed Ray Distance Functions (SRDF)
    Zins, Pierre
    Xu, Yuanlu
    Boyer, Edmond
    Wuhrer, Stefanie
    Tung, Tony
    2023 IEEE/CVF CONFERENCE ON COMPUTER VISION AND PATTERN RECOGNITION (CVPR), 2023, : 16696 - 16706
  • [3] COUPLED SIGNED-DISTANCE FUNCTIONS FOR IMPLICIT SURFACE RECONSTRUCTION
    Gong, Yuanhao
    Paul, Gregory
    Sbalzarini, Ivo F.
    2012 9TH IEEE INTERNATIONAL SYMPOSIUM ON BIOMEDICAL IMAGING (ISBI), 2012, : 1000 - 1003
  • [4] An adaptive descriptor for uncalibrated omnidirectional images - towards scene reconstruction by trifocal tensor
    Liu, Ming
    Alper, Bekir Tufan
    Siegwart, Roland
    2013 IEEE INTERNATIONAL CONFERENCE ON ROBOTICS AND AUTOMATION (ICRA), 2013, : 558 - 563
  • [5] Scene Reconstruction Using MRF Optimization with Image Content Adaptive Energy Functions
    Li, Ping
    Gunnewiek, Rene Klein
    de With, Peter H. N.
    ADVANCED CONCEPTS FOR INTELLIGENT VISION SYSTEMS, PROCEEDINGS, 2008, 5259 : 872 - +
  • [6] Pseudo-Plane Regularized Signed Distance Field for Neural Indoor Scene Reconstruction
    Li, Jing
    Yu, Jinpeng
    Wang, Ruoyu
    Gao, Shenghua
    INTERNATIONAL JOURNAL OF COMPUTER VISION, 2024,
  • [7] Localization of Transparent Objects Using Signed Distance Functions
    Yan, Jun
    Xia, Qing
    Ji, Yijun
    Zhang, Zhijiang
    SIXTH INTERNATIONAL CONFERENCE ON OPTICAL AND PHOTONIC ENGINEERING (ICOPEN 2018), 2018, 10827
  • [8] Diffusion generated motion using signed distance functions
    Esedoglu, Selim
    Ruuth, Steven
    Tsai, Richard
    JOURNAL OF COMPUTATIONAL PHYSICS, 2010, 229 (04) : 1017 - 1042
  • [9] Directed Ray Distance Functions for 3D Scene Reconstruction
    Kulkarni, Nilesh
    Johnson, Justin
    Fouhey, David F.
    COMPUTER VISION - ECCV 2022, PT II, 2022, 13662 : 201 - 219
  • [10] PoseSDF: Simultaneous 3D Human Shape Reconstruction and Gait Pose Estimation Using Signed Distance Functions
    Yang, Jianxin
    Liu, Yuxuan
    Gu, Xiao
    Yang, Guang-Zhong
    Guo, Yao
    2022 IEEE INTERNATIONAL CONFERENCE ON ROBOTICS AND AUTOMATION (ICRA 2022), 2022,