Exploring 3D-aware Latent Spaces for Efficiently Learning Numerous Scenes - Archive ouverte HAL Accéder directement au contenu
Pré-Publication, Document De Travail Année : 2024

Exploring 3D-aware Latent Spaces for Efficiently Learning Numerous Scenes

Résumé

We present a method enabling the scaling of NeRFs to learn a large number of semantically-similar scenes. We combine two techniques to improve the required training time and memory cost per scene. First, we learn a 3D-aware latent space in which we train Tri-Plane scene representations, hence reducing the resolution at which scenes are learned. Moreover, we present a way to share common information across scenes, hence allowing for a reduction of model complexity to learn a particular scene. Our method reduces effective per-scene memory costs by 44% and per-scene time costs by 86% when training 1000 scenes. Our project page can be found at https://3da-ae.github.io .
Fichier principal
Vignette du fichier
3da-ae.pdf (1.52 Mo) Télécharger le fichier
Origine : Fichiers produits par l'(les) auteur(s)
Licence : CC BY - Paternité

Dates et versions

hal-04509207 , version 1 (18-03-2024)

Licence

Paternité

Identifiants

  • HAL Id : hal-04509207 , version 1

Citer

Antoine Schnepf, Karim Kassab, Jean-Yves Franceschi, Laurent Caraffa, Flavian Vasile, et al.. Exploring 3D-aware Latent Spaces for Efficiently Learning Numerous Scenes. 2024. ⟨hal-04509207⟩
2 Consultations
3 Téléchargements

Partager

Gmail Facebook X LinkedIn More