StereoGS: Sparse-View 3D Gaussian Splatting via Stereo Priors
2026-06-29 • Computer Vision and Pattern Recognition
Computer Vision and Pattern Recognition
AI summaryⓘ
The authors address the problem of creating 3D scenes from very few pictures using 3D Gaussian Splatting, which usually struggles to build accurate shapes when there are limited views. They propose StereoGS, a method that uses stereo vision principles to better understand depth and scale by simulating pairs of images and enforcing consistency between them. Their technique also removes unnecessary parts of the 3D model and uses multi-view depth estimates to place details more accurately. Tests show StereoGS works better than previous methods on standard datasets, without making the system slower when viewing new angles.
3D Gaussian SplattingNovel View SynthesisSparse-view ReconstructionStereo VisionDepth RegularizationScale AmbiguityOpacity DecayMulti-view Depth EstimationBinocular Consistency3D Scene Reconstruction
Authors
Wenhao Yuan, Yiyuan Ge, Deli Cai
Abstract
3D Gaussian Splatting (3DGS) has achieved remarkable success in real-time novel view synthesis, yet it suffers from severe overfitting under sparse-view settings due to insufficient geometric constraints. While recent methods introduce monocular depth priors to mitigate this, they inherently struggle with scale ambiguity and cross-view inconsistency, leading to defective geometry. In this paper, we propose StereoGS, a novel sparse-view 3DGS framework that integrates stereo priors to establish reliable binocular consistency. Unlike scale-agnostic monocular constraints, StereoGS introduces a Stereo Depth Regularization by constructing virtual stereo pairs during optimization and leveraging a foundation stereo model to enforce absolute scale and binocular-consistent structures. To further suppress overfitting and eliminate redundant primitives, we design a Gradient-Aware Opacity Decay strategy that dynamically penalizes Gaussians based on their relative opacity gradient magnitudes. Combined with a Consistency-Aware Dense Initialization using zero-shot multi-view depth estimation, StereoGS effectively anchors primitives to accurate scene surfaces. Extensive experiments on LLFF, DTU, Mip-NeRF360, and Blender datasets demonstrate that StereoGS achieves state-of-the-art performance in sparse-view settings without incurring any additional inference overhead. Project Page: https://stringerywh00.github.io/StereoGS_project_page/