Search papers, labs, and topics across Lattice.
This paper introduces StereoGS, a novel framework for sparse-view 3D Gaussian Splatting that leverages stereo priors to enhance geometric consistency and mitigate overfitting. By employing a Stereo Depth Regularization technique that constructs virtual stereo pairs and utilizes a foundation stereo model, StereoGS addresses scale ambiguity and cross-view inconsistency prevalent in existing monocular depth methods. Extensive evaluations on multiple datasets reveal that StereoGS achieves state-of-the-art performance in sparse-view scenarios while maintaining efficient inference times.
StereoGS achieves unprecedented accuracy in sparse-view 3D Gaussian Splatting by enforcing binocular consistency, overcoming the limitations of monocular depth priors.
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/
