DUDF: Differentiable Unsigned Distance Fields with Hyperbolic Scaling

Mostrar todas las versiones(3)

In recent years, there has been a growing interest in training Neural Networks to approximate Unsigned Distance Fields (UDFs) for representing open surfaces in the context of 3D reconstruction. However, UDFs are non-differentiable at the zero level set which leads to significant errors in distances...

Descripción completa

Guardado en:
Detalles Bibliográficos
Autores principales: Fainstein, Miguel, Siless, Viviana, Iarussi, Emmanuel
Formato: Objeto de conferencia
Lenguaje:Inglés
Publicado: 2024
Materias:
Ciencias Informáticas
Neural implicit representations
Unsigned distance functions
Open surfaces
Deep neural networks
Acceso en línea:http://sedici.unlp.edu.ar/handle/10915/178813
Aporte de:
SEDICI (UNLP) de Universidad Nacional de La Plata
id I19-R120-10915-178813
record_format dspace
spelling I19-R120-10915-1788132025-05-08T20:07:53Z http://sedici.unlp.edu.ar/handle/10915/178813 DUDF: Differentiable Unsigned Distance Fields with Hyperbolic Scaling Fainstein, Miguel Siless, Viviana Iarussi, Emmanuel 2024-08 2024 2025-05-08T12:49:45Z en Ciencias Informáticas Neural implicit representations Unsigned distance functions Open surfaces Deep neural networks In recent years, there has been a growing interest in training Neural Networks to approximate Unsigned Distance Fields (UDFs) for representing open surfaces in the context of 3D reconstruction. However, UDFs are non-differentiable at the zero level set which leads to significant errors in distances and gradients, generally resulting in fragmented and discontinuous surfaces. In this paper, we propose to learn a hyperbolic scaling of the unsigned distance field, which defines a new Eikonal problem with distinct boundary conditions. This allows our formulation to integrate seamlessly with state-of-the-art continuously differentiable implicit neural representation networks, largely applied in the literature to represent signed distance fields. Our approach not only addresses the challenge of open surface representation but also demonstrates significant improvement in reconstruction quality and training performance. Moreover, the unlocked field’s differentiability allows the accurate computation of essential topological properties such as normal directions and curvatures, pervasive in downstream tasks such as rendering. Through extensive experiments, we validate our approach across various data sets and against competitive baselines. The results demonstrate enhanced accuracy and up to an order of magnitude increase in speed compared to previous methods. Sociedad Argentina de Informática e Investigación Operativa Objeto de conferencia Objeto de conferencia http://creativecommons.org/licenses/by-nc-sa/4.0/ Creative Commons Attribution-NonCommercial-ShareAlike 4.0 International (CC BY-NC-SA 4.0) application/pdf 113-113
institution Universidad Nacional de La Plata
institution_str I-19
repository_str R-120
collection SEDICI (UNLP)
language Inglés
topic Ciencias Informáticas
Neural implicit representations
Unsigned distance functions
Open surfaces
Deep neural networks
spellingShingle Ciencias Informáticas
Neural implicit representations
Unsigned distance functions
Open surfaces
Deep neural networks
Fainstein, Miguel
Siless, Viviana
Iarussi, Emmanuel
DUDF: Differentiable Unsigned Distance Fields with Hyperbolic Scaling
topic_facet Ciencias Informáticas
Neural implicit representations
Unsigned distance functions
Open surfaces
Deep neural networks
description In recent years, there has been a growing interest in training Neural Networks to approximate Unsigned Distance Fields (UDFs) for representing open surfaces in the context of 3D reconstruction. However, UDFs are non-differentiable at the zero level set which leads to significant errors in distances and gradients, generally resulting in fragmented and discontinuous surfaces. In this paper, we propose to learn a hyperbolic scaling of the unsigned distance field, which defines a new Eikonal problem with distinct boundary conditions. This allows our formulation to integrate seamlessly with state-of-the-art continuously differentiable implicit neural representation networks, largely applied in the literature to represent signed distance fields. Our approach not only addresses the challenge of open surface representation but also demonstrates significant improvement in reconstruction quality and training performance. Moreover, the unlocked field’s differentiability allows the accurate computation of essential topological properties such as normal directions and curvatures, pervasive in downstream tasks such as rendering. Through extensive experiments, we validate our approach across various data sets and against competitive baselines. The results demonstrate enhanced accuracy and up to an order of magnitude increase in speed compared to previous methods.
format Objeto de conferencia
Objeto de conferencia
author Fainstein, Miguel
Siless, Viviana
Iarussi, Emmanuel
author_facet Fainstein, Miguel
Siless, Viviana
Iarussi, Emmanuel
author_sort Fainstein, Miguel
title DUDF: Differentiable Unsigned Distance Fields with Hyperbolic Scaling
title_short DUDF: Differentiable Unsigned Distance Fields with Hyperbolic Scaling
title_full DUDF: Differentiable Unsigned Distance Fields with Hyperbolic Scaling
title_fullStr DUDF: Differentiable Unsigned Distance Fields with Hyperbolic Scaling
title_full_unstemmed DUDF: Differentiable Unsigned Distance Fields with Hyperbolic Scaling
title_sort dudf: differentiable unsigned distance fields with hyperbolic scaling
publishDate 2024
url http://sedici.unlp.edu.ar/handle/10915/178813
work_keys_str_mv AT fainsteinmiguel dudfdifferentiableunsigneddistancefieldswithhyperbolicscaling
AT silessviviana dudfdifferentiableunsigneddistancefieldswithhyperbolicscaling
AT iarussiemmanuel dudfdifferentiableunsigneddistancefieldswithhyperbolicscaling
_version_ 1847925378440495104

Ejemplares similares