High-Fidelity and Generalizable Neural Surface Reconstruction with Sparse Feature Volumes

Fuente: arXiv
Salvato in:
Dettagli Bibliografici
Autori principali: Fan, Aoxiang, Dumery, Corentin, Talabot, Nicolas, Le, Hieu, Fua, Pascal
Natura: Preprint
Pubblicazione: 2025
Soggetti:
Accesso online:
Tags: Aggiungi Tag
Nessun Tag, puoi essere il primo ad aggiungerne!!
_version_ 1866911045468028928
author Fan, Aoxiang
Dumery, Corentin
Talabot, Nicolas
Le, Hieu
Fua, Pascal
author_facet Fan, Aoxiang
Dumery, Corentin
Talabot, Nicolas
Le, Hieu
Fua, Pascal
contents Generalizable neural surface reconstruction has become a compelling technique to reconstruct from few images without per-scene optimization, where dense 3D feature volume has proven effective as a global representation of scenes. However, the dense representation does not scale well to increasing voxel resolutions, severely limiting the reconstruction quality. We thus present a sparse representation method, that maximizes memory efficiency and enables significantly higher resolution reconstructions on standard hardware. We implement this through a two-stage approach: First training a network to predict voxel occupancies from posed images and associated depth maps, then computing features and performing volume rendering only in voxels with sufficiently high occupancy estimates. To support this sparse representation, we developed custom algorithms for efficient sampling, feature aggregation, and querying from sparse volumes-overcoming the dense-volume assumptions inherent in existing works. Experiments on public datasets demonstrate that our approach reduces storage requirements by more than 50 times without performance degradation, enabling reconstructions at $512^3$ resolution compared to the typical $128^3$ on similar hardware, and achieving superior reconstruction accuracy over current state-of-the-art methods.
format Preprint
id arxiv_https___arxiv_org_abs_2507_05952
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle High-Fidelity and Generalizable Neural Surface Reconstruction with Sparse Feature Volumes
Fan, Aoxiang
Dumery, Corentin
Talabot, Nicolas
Le, Hieu
Fua, Pascal
Computer Vision and Pattern Recognition
Generalizable neural surface reconstruction has become a compelling technique to reconstruct from few images without per-scene optimization, where dense 3D feature volume has proven effective as a global representation of scenes. However, the dense representation does not scale well to increasing voxel resolutions, severely limiting the reconstruction quality. We thus present a sparse representation method, that maximizes memory efficiency and enables significantly higher resolution reconstructions on standard hardware. We implement this through a two-stage approach: First training a network to predict voxel occupancies from posed images and associated depth maps, then computing features and performing volume rendering only in voxels with sufficiently high occupancy estimates. To support this sparse representation, we developed custom algorithms for efficient sampling, feature aggregation, and querying from sparse volumes-overcoming the dense-volume assumptions inherent in existing works. Experiments on public datasets demonstrate that our approach reduces storage requirements by more than 50 times without performance degradation, enabling reconstructions at $512^3$ resolution compared to the typical $128^3$ on similar hardware, and achieving superior reconstruction accuracy over current state-of-the-art methods.
title High-Fidelity and Generalizable Neural Surface Reconstruction with Sparse Feature Volumes
topic Computer Vision and Pattern Recognition
url https://arxiv.org/abs/2507.05952