Neural Vector Fields for Surface Representation and Inference

Rella, Edoardo Mello; Chhatkuli, Ajad; Konukoğlu, Ender; Gool, Luc Van

doi:10.48550/arxiv.2204.06552

Cited by 1 publication

(1 citation statement)

References 0 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…signed distance fields [JBS06], generally represent the surface as a level set of a scalar field, f : R 3 → R, and this limits the methods' applicability to watertight surfaces. Venkatesh et al [VKS*21] adopt an approach in which they return the closest point on the surface for a given query point, whereas Rella et al [RCKV22] find a vector field and extract the surface as the set of points where the vector field is zero. Palmer et al…”

Section: Related Workmentioning

confidence: 99%

Neural Representation of Open Surfaces

Christiansen

Bærentzen

Paulsen

et al. 2023

Computer Graphics Forum

View full text Add to dashboard Cite

Neural implicit surfaces have emerged as an effective, learnable representation for shapes of arbitrary topology. However, representing open surfaces remains a challenge. Different methods, such as unsigned distance fields (UDF), have been proposed to tackle this issue, but a general solution remains elusive. The generalized winding number (GWN), which is often used to distinguish interior points from exterior points of 3D shapes, is arguably the most promising approach. The GWN changes smoothly in regions where there is a hole in the surface, but it is discontinuous at points on the surface. Effectively, this means that it can be used in lieu of an implicit surface representation while providing information about holes, but, unfortunately, it does not provide information about the distance to the surface necessary for e.g. ray tracing, and special care must be taken when implementing surface reconstruction. Therefore, we introduce the semi‐signed distance field (SSDF) representation which comprises both the GWN and the surface distance. We compare the GWN and SSDF representations for the applications of surface reconstruction, interpolation, reconstruction from partial data, and latent vector analysis using two very different data sets. We find that both the GWN and SSDF are well suited for neural representation of open surfaces.

show abstract

Section: Related Workmentioning

confidence: 99%