Guardado en:
Detalles Bibliográficos
Autores principales: Sundararaman, Ramana, Donati, Nicolas, Melzi, Simone, Corman, Etienne, Ovsjanikov, Maks
Formato: Preprint
Publicado: 2024
Materias:
Acceso en línea:https://arxiv.org/abs/2410.08225
Etiquetas: Agregar Etiqueta
Sin Etiquetas, Sea el primero en etiquetar este registro!
_version_ 1866917800135622656
author Sundararaman, Ramana
Donati, Nicolas
Melzi, Simone
Corman, Etienne
Ovsjanikov, Maks
author_facet Sundararaman, Ramana
Donati, Nicolas
Melzi, Simone
Corman, Etienne
Ovsjanikov, Maks
contents We introduce a novel data-driven approach aimed at designing high-quality shape deformations based on a coarse localized input signal. Unlike previous data-driven methods that require a global shape encoding, we observe that detail-preserving deformations can be estimated reliably without any global context in certain scenarios. Building on this intuition, we leverage Jacobians defined in a one-ring neighborhood as a coarse representation of the deformation. Using this as the input to our neural network, we apply a series of MLPs combined with feature smoothing to learn the Jacobian corresponding to the detail-preserving deformation, from which the embedding is recovered by the standard Poisson solve. Crucially, by removing the dependence on a global encoding, every \textit{point} becomes a training example, making the supervision particularly lightweight. Moreover, when trained on a class of shapes, our approach demonstrates remarkable generalization across different object categories. Equipped with this novel network, we explore three main tasks: refining an approximate shape correspondence, unsupervised deformation and mapping, and shape editing. Our code is made available at https://github.com/sentient07/LJN
format Preprint
id arxiv_https___arxiv_org_abs_2410_08225
institution arXiv
publishDate 2024
record_format arxiv
spellingShingle Deformation Recovery: Localized Learning for Detail-Preserving Deformations
Sundararaman, Ramana
Donati, Nicolas
Melzi, Simone
Corman, Etienne
Ovsjanikov, Maks
Graphics
We introduce a novel data-driven approach aimed at designing high-quality shape deformations based on a coarse localized input signal. Unlike previous data-driven methods that require a global shape encoding, we observe that detail-preserving deformations can be estimated reliably without any global context in certain scenarios. Building on this intuition, we leverage Jacobians defined in a one-ring neighborhood as a coarse representation of the deformation. Using this as the input to our neural network, we apply a series of MLPs combined with feature smoothing to learn the Jacobian corresponding to the detail-preserving deformation, from which the embedding is recovered by the standard Poisson solve. Crucially, by removing the dependence on a global encoding, every \textit{point} becomes a training example, making the supervision particularly lightweight. Moreover, when trained on a class of shapes, our approach demonstrates remarkable generalization across different object categories. Equipped with this novel network, we explore three main tasks: refining an approximate shape correspondence, unsupervised deformation and mapping, and shape editing. Our code is made available at https://github.com/sentient07/LJN
title Deformation Recovery: Localized Learning for Detail-Preserving Deformations
topic Graphics
url https://arxiv.org/abs/2410.08225