Centre de documentation
scientifique

DiffusionNet: discretization agnostic learning on surfaces / Nicholas Sharp in ACM Transactions on Graphics, TOG, Vol 41 n° 3 (June 2022)  

Public [article]  inACM Transactions on Graphics, TOG > Vol 41 n° 3 (June 2022) . - pp 1 - 16 Titre : DiffusionNet: discretization agnostic learning on surfaces Type de document : Article/Communication Auteurs : Nicholas Sharp, Auteur ; Souhaib Attaiki, Auteur ; K. Crane, Auteur ; et al., Auteur Année de publication : 2022 Article en page(s) : pp 1 - 16 Note générale : bibliographie Langues : Anglais (eng) Descripteur : [Vedettes matières IGN] Intelligence artificielle [Termes IGN] apprentissage profond [Termes IGN] discrétisation [Termes IGN] maillage [Termes IGN] Perceptron multicouche [Termes IGN] semis de points [Termes IGN] Triangular Regular Network [Termes IGN] voisinage (relation topologique) Résumé : (auteur) 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. Numéro de notice : A2022-321 Affiliation des auteurs : non IGN Thématique : IMAGERIE/INFORMATIQUE Nature : Article DOI : 10.1145/3507905 Date de publication en ligne : 07/03/2022 En ligne : https://doi.org/10.1145/3507905 Format de la ressource électronique : URL article Permalink : https://documentation.ensg.eu/index.php?lvl=notice_display&id=100369 [article]