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

Given only a set of images, neural implicit surface representation has shownits capability in 3D surface reconstruction. However, as the nature ofper-scene optimization is based on the volumetric rendering of color, previousneural implicit surface reconstruction methods usually fail in low-texturedregions, including the floors, walls, etc., which commonly exist for indoorscenes. Being aware of the fact that these low-textured regions usuallycorrespond to planes, without introducing additional ground-truth supervisorysignals or making additional assumptions about the room layout, we propose toleverage a novel Pseudo Plane-regularized Signed Distance Field (P$^2$SDF) forindoor scene reconstruction. Specifically, we consider adjacent pixels withsimilar colors to be on the same pseudo planes. The plane parameters are thenestimated on the fly during training by an efficient and effective two-stepscheme. Then the signed distances of the points on the planes are regularizedby the estimated plane parameters in the training phase. As the unsupervisedplane segments are usually noisy and inaccurate, we propose to assign differentweights to the sampled points on the plane in plane estimation as well as theregularization loss. The weights come by fusing the plane segments fromdifferent views. As the sampled rays in the planar regions are redundant,leading to inefficient training, we further propose a keypoint-guided rayssampling strategy that attends to the informative textured regions with largecolor variations, and the implicit network gets a better reconstruction,compared with the original uniform ray sampling strategy. Experiments show thatour P$^2$SDF achieves competitive reconstruction performance in Manhattanscenes. Further, as we do not introduce any additional room layout assumption,our P$^2$SDF generalizes well to the reconstruction of non-Manhattan scenes.

P$^2$SDF for Neural Indoor Scene Reconstruction