Share your thoughts, 1 month free Claude Pro on usSee more
WorkDL logo mark

PairDropGS: Paired Dropout-Induced Consistency Regularization for Sparse-View Gaussian Splatting

About

Dropout-based sparse-view 3D Gaussian Splatting (3DGS) methods alleviate overfitting by randomly suppressing Gaussian primitives during training. Existing methods mainly focus on designing increasingly sophisticated dropout strategies, while they overlook the resulting inconsistencies among different dropped Gaussian subsets. This oversight often leads to unstable reconstruction and suboptimal Gaussian representation learning.In this paper, we revisit dropout-based sparse-view 3DGS from a consistency regularization perspective and propose PairDropGS, a Paired Dropout-induced Consistency Regularization framework for sparse-view Gaussian splatting. Specifically, PairDropGS first constructs a pair of the dropped Gaussian subsets from a shared Gaussian field and designs a low-frequency consistency regularization to constrain their low-frequency rendered structures. This design encourages the shared Gaussian field to preserve stable scene layout and coarse geometry under different random dropouts, while avoiding excessive constraints on ambiguous high-frequency details. Moreover, we introduce a progressive consistency scheduling strategy to gradually strengthen the consistency regularization during training for stability and robustness of reconstruction. Extensive experiments on widely-used sparse-view benchmarks demonstrate that PairDropGS achieves superior training stability, significantly outperforms existing dropout-based 3DGS methods in reconstruction quality, while exhibiting the simplicity and plug-and-play nature for improving dropout-based optimization.

Hantang Li, Qiang Zhu, Xiandong Meng, Xingtao Wang, Debin Zhao, Xiaopeng Fan• 2026

Related benchmarks

TaskDatasetResultRank
Novel View SynthesisLLFF 3-view
PSNR20.89
162
Novel View SynthesisLLFF 6-view
PSNR24.93
129
Novel View SynthesisLLFF 9-view
PSNR26.31
121
Novel View SynthesisMip-NeRF 360 12-view
PSNR20.12
53
Novel View SynthesisMip-NeRF 360 24-view
PSNR24.21
38
Novel View SynthesisBlender standard sparse-view 8-view
PSNR25.57
9
Showing 6 of 6 rows

Other info

Follow for update