We gratefully acknowledge support from
the Simons Foundation and member institutions.
Full-text links:

Download:

Current browse context:

cs.CV

Change to browse by:

cs

References & Citations

DBLP - CS Bibliography

Bookmark

(what is this?)
CiteULike logo BibSonomy logo Mendeley logo del.icio.us logo Digg logo Reddit logo ScienceWISE logo

Computer Science > Computer Vision and Pattern Recognition

Title: Dynamic View Synthesis from Dynamic Monocular Video

Abstract: We present an algorithm for generating novel views at arbitrary viewpoints and any input time step given a monocular video of a dynamic scene. Our work builds upon recent advances in neural implicit representation and uses continuous and differentiable functions for modeling the time-varying structure and the appearance of the scene. We jointly train a time-invariant static NeRF and a time-varying dynamic NeRF, and learn how to blend the results in an unsupervised manner. However, learning this implicit function from a single video is highly ill-posed (with infinitely many solutions that match the input video). To resolve the ambiguity, we introduce regularization losses to encourage a more physically plausible solution. We show extensive quantitative and qualitative results of dynamic view synthesis from casually captured videos.
Comments: Project webpage: this https URL
Subjects: Computer Vision and Pattern Recognition (cs.CV)
Cite as: arXiv:2105.06468 [cs.CV]
  (or arXiv:2105.06468v1 [cs.CV] for this version)

Submission history

From: Chen Gao [view email]
[v1] Thu, 13 May 2021 17:59:50 GMT (29708kb,D)

Link back to: arXiv, form interface, contact.