SparseCraft: Few-Shot Neural Reconstruction through Stereopsis Guided Geometric Linearization

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Younes, Mae, Ouasfi, Amine, Boukhayma, Adnane
Format: Preprint
Published: 2024
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866929427575734272
author Younes, Mae
Ouasfi, Amine
Boukhayma, Adnane
author_facet Younes, Mae
Ouasfi, Amine
Boukhayma, Adnane
contents We present a novel approach for recovering 3D shape and view dependent appearance from a few colored images, enabling efficient 3D reconstruction and novel view synthesis. Our method learns an implicit neural representation in the form of a Signed Distance Function (SDF) and a radiance field. The model is trained progressively through ray marching enabled volumetric rendering, and regularized with learning-free multi-view stereo (MVS) cues. Key to our contribution is a novel implicit neural shape function learning strategy that encourages our SDF field to be as linear as possible near the level-set, hence robustifying the training against noise emanating from the supervision and regularization signals. Without using any pretrained priors, our method, called SparseCraft, achieves state-of-the-art performances both in novel-view synthesis and reconstruction from sparse views in standard benchmarks, while requiring less than 10 minutes for training.
format Preprint
id arxiv_https___arxiv_org_abs_2407_14257
institution arXiv
publishDate 2024
record_format arxiv
spellingShingle SparseCraft: Few-Shot Neural Reconstruction through Stereopsis Guided Geometric Linearization
Younes, Mae
Ouasfi, Amine
Boukhayma, Adnane
Computer Vision and Pattern Recognition
We present a novel approach for recovering 3D shape and view dependent appearance from a few colored images, enabling efficient 3D reconstruction and novel view synthesis. Our method learns an implicit neural representation in the form of a Signed Distance Function (SDF) and a radiance field. The model is trained progressively through ray marching enabled volumetric rendering, and regularized with learning-free multi-view stereo (MVS) cues. Key to our contribution is a novel implicit neural shape function learning strategy that encourages our SDF field to be as linear as possible near the level-set, hence robustifying the training against noise emanating from the supervision and regularization signals. Without using any pretrained priors, our method, called SparseCraft, achieves state-of-the-art performances both in novel-view synthesis and reconstruction from sparse views in standard benchmarks, while requiring less than 10 minutes for training.
title SparseCraft: Few-Shot Neural Reconstruction through Stereopsis Guided Geometric Linearization
topic Computer Vision and Pattern Recognition
url https://arxiv.org/abs/2407.14257