Reproduce “DeVRF: Fast Deformable Voxel Radiance Fields for Dynamic Scenes” on OpenResearch
Turn your favorite coding agents into research agents that review literature, develop hypotheses, and run experiments.
Universal app for Apple silicon and Intel. Requires macOS 11 or later.
About this paper
Jia-Wei Liu, Yan-Pei Cao, Weijia Mao, Wenqiao Zhang, David Junhao Zhang, Jussi Keppo, Ying Shan, Xiaohu Qie and 1 more
DeVRF learns dynamic scenes in two stages: it fits static geometry and appearance with voxel grids from many views, then learns time-varying motion from a few fixed-camera videos using a 4D voxel deformation field. On tested scenes, it trains in roughly 10 minutes with comparable image quality and about a 100× speedup, but may struggle with drastic deformation.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.