Signed Distance Function Computation from an Implicit Surface
This addresses a specific computational geometry problem for researchers or practitioners needing accurate SDFs from implicit surfaces, but it appears incremental as it builds on existing neural network methods.
The paper tackles the problem of converting an implicit surface into a Signed Distance Function (SDF) while exactly preserving the zero level-set, and the result is a technique that embeds the input implicit in a neural network trained to minimize an SDF loss function.
We describe in this short note a technique to convert an implicit surface into a Signed Distance Function (SDF) while exactly preserving the zero level-set of the implicit. The proposed approach relies on embedding the input implicit in the final layer of a neural network, which is trained to minimize a loss function characterizing the SDF.