Geometric prior-guided self-supervised learning for multi-view stereo

HIGHLIGHTS

  • who: Fenghao Zhang and colleagues from the School of Biomedical Engineering, South-Central Minzu University, Wuhan, China have published the research work: Geometric Prior-Guided Self-Supervised Learning for Multi-View Stereo, in the Journal: (JOURNAL)
  • what: The authors propose the geometric prior-guided multi-view stereo (GP-MVS) approach for self-supervised learning. Specifically, the authors propose two types of depth pseudo-labels, sparse and semi-dense, based on the geometry information of the 3D scene. Sparse Pseudo-Label Semi-Dense Pseudo-label SfM Input Images Depth Estimation Filtering Initial Depth Map Sparse Point . . .

     

    Logo ScioWire Beta black

    If you want to have access to all the content you need to log in!

    Thanks :)

    If you don't have an account, you can create one here.

     

Scroll to Top

Add A Knowledge Base Question !

+ = Verify Human or Spambot ?