SurfelSplat: Learning Efficient and Generalizable Gaussian Surfel Representations for Sparse-View Surface Reconstruction

TL;DR

SurfelSplat introduces a fast, feed-forward approach for sparse-view 3D surface reconstruction using Gaussian surfel representations, leveraging cross-view feature aggregation guided by Nyquist sampling to achieve high accuracy and efficiency.

Contribution

It proposes a novel, efficient framework that generates accurate Gaussian surfel representations from sparse views without per-scene optimization, utilizing a Nyquist sampling-based feature aggregation.

Findings

Achieves comparable reconstruction quality to state-of-the-art methods.

Predicts Gaussian surfels within 1 second, 100x faster than optimization-based methods.

Demonstrates effectiveness on DTU benchmarks with sparse input views.

Abstract

3D Gaussian Splatting (3DGS) has demonstrated impressive performance in 3D scene reconstruction. Beyond novel view synthesis, it shows great potential for multi-view surface reconstruction. Existing methods employ optimization-based reconstruction pipelines that achieve precise and complete surface extractions. However, these approaches typically require dense input views and high time consumption for per-scene optimization. To address these limitations, we propose SurfelSplat, a feed-forward framework that generates efficient and generalizable pixel-aligned Gaussian surfel representations from sparse-view images. We observe that conventional feed-forward structures struggle to recover accurate geometric attributes of Gaussian surfels because the spatial frequency of pixel-aligned primitives exceeds Nyquist sampling rates. Therefore, we propose a cross-view feature aggregation module based on the Nyquist sampling theorem. Specifically, we first adapt the geometric forms of Gaussian surfels with spatial sampling rate-guided low-pass filters. We then project the filtered surfels across all input views to obtain cross-view feature correlations. By processing these correlations through a specially designed feature fusion network, we can finally regress Gaussian surfels with precise geometry. Extensive experiments on DTU reconstruction benchmarks demonstrate that our model achieves comparable results with state-of-the-art methods, and predict Gaussian surfels within 1 second, offering a 100x speedup without costly per-scene training.