Multi-View Stereo Network with Regional Consistency and Discrepancy Cost Volume
DOI:
Author:
Affiliation:

Tianjin University of Technology

Clc Number:

Fund Project:

National Natural Science Foundation of China (No. 62302335)

  • Article
  • |
  • Figures
  • |
  • Metrics
  • |
  • Reference
  • |
  • Related
  • |
  • Cited by
  • |
  • Materials
  • |
  • Comments
    Abstract:

    The goal of Multi-View Stereo (MVS) is to robustly recover an accurate 3D point cloud from multiple views. In this pa-per, we propose a novel Multi-View Stereo network with Regional consistency and Discrepancy cost volume, denoted as MVSRD. Firstly, a full-feature interaction transformer is presented, which learns the regional consistency between the reference and source views, improving the robustness of reconstruction. Secondly, a discrepancy cost volume is designed to emphasize pixel-level differences between feature volumes. It facilitates the constructing of high-quality cost volume to enhance the accuracy of reconstruction. Extensive experiments on the DTU and Tanks & Temples datasets demonstrate that our MVSRD achieves state-of-the-art performance.

    Reference
    Related
    Cited by
Get Citation
Share
Article Metrics
  • Abstract:
  • PDF:
  • HTML:
  • Cited by:
History
  • Received:June 08,2024
  • Revised:August 09,2024
  • Adopted:September 04,2024
  • Online:
  • Published: