HotSpot: Signed Distance Function Optimization with an Asymptotically Sufficient Condition
This work addresses the problem of accurate and stable neural signed distance function optimization for researchers in computer vision and graphics, offering a novel loss function that improves reconstruction quality, though it is incremental in advancing existing methods.
The authors tackled the problem of optimizing neural signed distance functions by addressing the insufficiency of existing losses like the eikonal loss, which cannot guarantee true distance functions and suffer from stability issues. They proposed HotSpot, a method using a loss based on a screened Poisson equation, which provides an asymptotically sufficient condition for convergence to true distance functions, leading to better surface reconstruction and more accurate distance approximations in experiments on 2D and 3D datasets.
We propose a method, HotSpot, for optimizing neural signed distance functions. Existing losses, such as the eikonal loss, act as necessary but insufficient constraints and cannot guarantee that the recovered implicit function represents a true distance function, even if the output minimizes these losses almost everywhere. Furthermore, the eikonal loss suffers from stability issues in optimization. Finally, in conventional methods, regularization losses that penalize surface area distort the reconstructed signed distance function. We address these challenges by designing a loss function using the solution of a screened Poisson equation. Our loss, when minimized, provides an asymptotically sufficient condition to ensure the output converges to a true distance function. Our loss also leads to stable optimization and naturally penalizes large surface areas. We present theoretical analysis and experiments on both challenging 2D and 3D datasets and show that our method provides better surface reconstruction and a more accurate distance approximation.