Reproduce “Multi-NeuS: 3D Head Portraits from Single Image with Neural Implicit Functions” 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
Egor Burkov, Ruslan Rakhimov, Aleksandr Safin, Evgeny Burnaev, Victor Lempitsky
Multi-NeuS reconstructs textured 3D heads from one or a few images by training a NeuS-based model across people, sharing layers, then adapting person-specific layers before fine-tuning the whole model. On H3DS, its one-view mean errors were 1.84 mm for faces and 12.74 mm for heads, close to H3D-Net’s one-view results; unseen backs remain unreliable.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.