Epipolar-Free 3D Gaussian Splatting for Generalizable Novel View Synthesis

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Min, Zhiyuan, Luo, Yawei, Sun, Jianwen, Yang, Yi
Format: Preprint
Published: 2024
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866910678765273088
author Min, Zhiyuan
Luo, Yawei
Sun, Jianwen
Yang, Yi
author_facet Min, Zhiyuan
Luo, Yawei
Sun, Jianwen
Yang, Yi
contents Generalizable 3D Gaussian splitting (3DGS) can reconstruct new scenes from sparse-view observations in a feed-forward inference manner, eliminating the need for scene-specific retraining required in conventional 3DGS. However, existing methods rely heavily on epipolar priors, which can be unreliable in complex realworld scenes, particularly in non-overlapping and occluded regions. In this paper, we propose eFreeSplat, an efficient feed-forward 3DGS-based model for generalizable novel view synthesis that operates independently of epipolar line constraints. To enhance multiview feature extraction with 3D perception, we employ a selfsupervised Vision Transformer (ViT) with cross-view completion pre-training on large-scale datasets. Additionally, we introduce an Iterative Cross-view Gaussians Alignment method to ensure consistent depth scales across different views. Our eFreeSplat represents an innovative approach for generalizable novel view synthesis. Different from the existing pure geometry-free methods, eFreeSplat focuses more on achieving epipolar-free feature matching and encoding by providing 3D priors through cross-view pretraining. We evaluate eFreeSplat on wide-baseline novel view synthesis tasks using the RealEstate10K and ACID datasets. Extensive experiments demonstrate that eFreeSplat surpasses state-of-the-art baselines that rely on epipolar priors, achieving superior geometry reconstruction and novel view synthesis quality. Project page: https://tatakai1.github.io/efreesplat/.
format Preprint
id arxiv_https___arxiv_org_abs_2410_22817
institution arXiv
publishDate 2024
record_format arxiv
spellingShingle Epipolar-Free 3D Gaussian Splatting for Generalizable Novel View Synthesis
Min, Zhiyuan
Luo, Yawei
Sun, Jianwen
Yang, Yi
Computer Vision and Pattern Recognition
Generalizable 3D Gaussian splitting (3DGS) can reconstruct new scenes from sparse-view observations in a feed-forward inference manner, eliminating the need for scene-specific retraining required in conventional 3DGS. However, existing methods rely heavily on epipolar priors, which can be unreliable in complex realworld scenes, particularly in non-overlapping and occluded regions. In this paper, we propose eFreeSplat, an efficient feed-forward 3DGS-based model for generalizable novel view synthesis that operates independently of epipolar line constraints. To enhance multiview feature extraction with 3D perception, we employ a selfsupervised Vision Transformer (ViT) with cross-view completion pre-training on large-scale datasets. Additionally, we introduce an Iterative Cross-view Gaussians Alignment method to ensure consistent depth scales across different views. Our eFreeSplat represents an innovative approach for generalizable novel view synthesis. Different from the existing pure geometry-free methods, eFreeSplat focuses more on achieving epipolar-free feature matching and encoding by providing 3D priors through cross-view pretraining. We evaluate eFreeSplat on wide-baseline novel view synthesis tasks using the RealEstate10K and ACID datasets. Extensive experiments demonstrate that eFreeSplat surpasses state-of-the-art baselines that rely on epipolar priors, achieving superior geometry reconstruction and novel view synthesis quality. Project page: https://tatakai1.github.io/efreesplat/.
title Epipolar-Free 3D Gaussian Splatting for Generalizable Novel View Synthesis
topic Computer Vision and Pattern Recognition
url https://arxiv.org/abs/2410.22817