2D Neural Fields with Learned Discontinuities
No Thumbnail Available
Date
2025
Journal Title
Journal ISSN
Volume Title
Publisher
The Eurographics Association and John Wiley & Sons Ltd.
Abstract
Effective representation of 2D images is fundamental in digital image processing, where traditional methods like raster and vector graphics struggle with sharpness and textural complexity, respectively. Current neural fields offer high fidelity and resolution independence but require predefined meshes with known discontinuities, restricting their utility. We observe that by treating all mesh edges as potential discontinuities, we can represent the discontinuity magnitudes as continuous variables and optimize. We further introduce a novel discontinuous neural field model that jointly approximates the target image and recovers discontinuities. Through systematic evaluations, our neural field outperforms other methods that fit unknown discontinuities with discontinuous representations, exceeding Field of Junction and Boundary Attention by over 11dB in both denoising and super-resolution tasks and achieving 3.5× smaller Chamfer distances than Mumford-Shah-based methods. It also surpasses InstantNGP with improvements of more than 5dB (denoising) and 10dB (super-resolution). Additionally, our approach shows remarkable capability in approximating complex artistic and natural images and cleaning up diffusion-generated depth maps.
Description
CCS Concepts: Computing methodologies → Image representations; Reconstruction; Neural networks
@article{10.1111:cgf.70023,
journal = {Computer Graphics Forum},
title = {{2D Neural Fields with Learned Discontinuities}},
author = {Liu, Chenxi and Wang, Siqi and Fisher, Matthew and Aneja, Deepali and Jacobson, Alec},
year = {2025},
publisher = {The Eurographics Association and John Wiley & Sons Ltd.},
ISSN = {1467-8659},
DOI = {10.1111/cgf.70023}
}