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 Tools annual

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 Tools monthly

Global: Zendy Plus monthly

Presents the access of premium content as premium feature

Premium Content

Global: Zendy Plus annual

Global: Zendy Free Plan

In recent years, neural implicit surface reconstruction methods have becomepopular for multi-view 3D reconstruction. In contrast to traditional multi-viewstereo methods, these approaches tend to produce smoother and more completereconstructions due to the inductive smoothness bias of neural networks.State-of-the-art neural implicit methods allow for high-quality reconstructionsof simple scenes from many input views. Yet, their performance dropssignificantly for larger and more complex scenes and scenes captured fromsparse viewpoints. This is caused primarily by the inherent ambiguity in theRGB reconstruction loss that does not provide enough constraints, in particularin less-observed and textureless areas. Motivated by recent advances in thearea of monocular geometry prediction, we systematically explore the utilitythese cues provide for improving neural implicit surface reconstruction. Wedemonstrate that depth and normal cues, predicted by general-purpose monocularestimators, significantly improve reconstruction quality and optimization time.Further, we analyse and investigate multiple design choices for representingneural implicit surfaces, ranging from monolithic MLP models over single-gridto multi-resolution grid representations. We observe that geometric monocularpriors improve performance both for small-scale single-object as well aslarge-scale multi-object scenes, independent of the choice of representation.

MonoSDF: Exploring Monocular Geometric Cues for Neural Implicit Surface Reconstruction