English (United Kingdom)

https://curated-unify.zendy.io/wp-json/zendy-region/v1/featured_content/oa?rat=en

https://curated-unify.zendy.io/wp-json/zendy-region/v1/highlighted_journal/

Global: Zendy Plus annual

Presents the access of premium content as premium feature

Premium Content

Presents the keyphrase highlighting as premium feature

Keyphrase Highlighting

Presents the summarisation as premium feature

Summarisation

Insights

Presents the pdf analysis as premium feature

PDF Analysis

Presents the zaia usage as premium feature

ZAIA

Global: Zendy Plus monthly

Global: Zendy Tools annual

Global: Zendy Tools monthly

Global: Zendy Free Plan

We present a novel neural surface reconstruction method, called NeuS, forreconstructing objects and scenes with high fidelity from 2D image inputs.Existing neural surface reconstruction approaches, such as DVR and IDR, requireforeground mask as supervision, easily get trapped in local minima, andtherefore struggle with the reconstruction of objects with severeself-occlusion or thin structures. Meanwhile, recent neural methods for novelview synthesis, such as NeRF and its variants, use volume rendering to producea neural scene representation with robustness of optimization, even for highlycomplex objects. However, extracting high-quality surfaces from this learnedimplicit representation is difficult because there are not sufficient surfaceconstraints in the representation. In NeuS, we propose to represent a surfaceas the zero-level set of a signed distance function (SDF) and develop a newvolume rendering method to train a neural SDF representation. We observe thatthe conventional volume rendering method causes inherent geometric errors (i.e.bias) for surface reconstruction, and therefore propose a new formulation thatis free of bias in the first order of approximation, thus leading to moreaccurate surface reconstruction even without the mask supervision. Experimentson the DTU dataset and the BlendedMVS dataset show that NeuS outperforms thestate-of-the-arts in high-quality surface reconstruction, especially forobjects and scenes with complex structures and self-occlusion.

NeuS: Learning Neural Implicit Surfaces by Volume Rendering for Multi-view Reconstruction