DiffusionNet: Discretization Agnostic Learning on Surfaces - Inria - Institut national de recherche en sciences et technologies du numérique
Journal Articles ACM Transactions on Graphics Year : 2022

DiffusionNet: Discretization Agnostic Learning on Surfaces

Abstract

We introduce a new general-purpose approach to deep learning on three-dimensional surfaces based on the insight that a simple diffusion layer is highly effective for spatial communication. The resulting networks are automatically robust to changes in resolution and sampling of a surface—a basic property that is crucial for practical applications. Our networks can be discretized on various geometric representations, such as triangle meshes or point clouds, and can even be trained on one representation and then applied to another. We optimize the spatial support of diffusion as a continuous network parameter ranging from purely local to totally global, removing the burden of manually choosing neighborhood sizes. The only other ingredients in the method are a multi-layer perceptron applied independently at each point and spatial gradient features to support directional filters. The resulting networks are simple, robust, and efficient. Here, we focus primarily on triangle mesh surfaces and demonstrate state-of-the-art results for a variety of tasks, including surface classification, segmentation, and non-rigid correspondence.
Fichier principal
Vignette du fichier
DiffusionNet_final_cmp.pdf (2.73 Mo) Télécharger le fichier
Origin Publisher files allowed on an open archive

Dates and versions

hal-03911881 , version 1 (23-12-2022)

Identifiers

Cite

Nicholas Sharp, Souhaib Attaiki, Keenan Crane, Maks Ovsjanikov. DiffusionNet: Discretization Agnostic Learning on Surfaces. ACM Transactions on Graphics, 2022, 41 (3), pp.1-16. ⟨10.1145/3507905⟩. ⟨hal-03911881⟩
122 View
121 Download

Altmetric

Share

More