Geometric implicit neural representations for signed distance functions
It addresses the problem of enhancing 3D reconstruction quality for applications in computer vision and graphics, but is incremental as it surveys existing methods rather than introducing new ones.
This survey reviews geometric implicit neural representations (INRs) for approximating signed distance functions (SDFs) in 3D surface reconstruction, highlighting advancements in using differential geometry tools like normals and curvatures in loss functions to improve accuracy.
\textit{Implicit neural representations} (INRs) have emerged as a promising framework for representing signals in low-dimensional spaces. This survey reviews the existing literature on the specialized INR problem of approximating \textit{signed distance functions} (SDFs) for surface scenes, using either oriented point clouds or a set of posed images. We refer to neural SDFs that incorporate differential geometry tools, such as normals and curvatures, in their loss functions as \textit{geometric} INRs. The key idea behind this 3D reconstruction approach is to include additional \textit{regularization} terms in the loss function, ensuring that the INR satisfies certain global properties that the function should hold -- such as having unit gradient in the case of SDFs. We explore key methodological components, including the definition of INR, the construction of geometric loss functions, and sampling schemes from a differential geometry perspective. Our review highlights the significant advancements enabled by geometric INRs in surface reconstruction from oriented point clouds and posed images.