PAGaS: Pixel-Aligned 1DoF Gaussian Splatting for Depth Refinement

arXiv cs.CV / 4/27/2026

📰 NewsDeveloper Stack & InfrastructureModels & Research

共有:

Key Points

The paper introduces Pixel-Aligned 1DoF Gaussian Splatting (PAGaS), adapting Gaussian Splatting from novel view synthesis to the multi-view stereo (MVS) depth refinement problem.
PAGaS models each pixel’s depth using one-degree-of-freedom (1DoF) Gaussians, constraining Gaussian positions and sizes via back-projected pixel volumes so that depth is the only parameter optimized.
By tightly constraining the representation during optimization, PAGaS aims to improve geometric fidelity that earlier GS variants struggled to capture accurately.
The authors report quantitative improvements on challenging 3D reconstruction benchmarks, evaluated against both geometric and learning-based MVS baselines.
The work includes released code via the project website, enabling further testing and adoption of PAGaS.

Abstract

Gaussian Splatting (GS) has emerged as an efficient approach for high-quality novel view synthesis. While early GS variants struggled to accurately model the scene's geometry, recent advancements constraining the Gaussians' spread and shapes, such as 2D Gaussian Splatting, have significantly improved geometric fidelity. In this paper, we present Pixel-Aligned 1DoF Gaussian Splatting (PAGaS) that adapts the GS representation from novel view synthesis to the multi-view stereo depth task. Our key contribution is modeling a pixel's depth using one-degree-of-freedom (1DoF) Gaussians that remain tightly constrained during optimization. Unlike existing approaches, our Gaussians' positions and sizes are restricted by the back-projected pixel volumes, leaving depth as the sole degree of freedom to optimize. PAGaS produces highly detailed depths, as illustrated in Figure 1. We quantitatively validate these improvements on top of reference geometric and learning-based multi-view stereo baselines on challenging 3D reconstruction benchmarks. Code: davidrecasens.github.io/pagas