DropoutGS: Dropping Out Gaussians for Better Sparse-view Rendering

2025-04-13

AI Summary

In sparse-view settings, 3D Gaussian Splatting (3DGS) suffers from severe overfitting, geometric distortion, and prominent artifacts. To address these issues, we propose two key innovations: (1) Random Gaussian Dropping Regularization (RDR), the first of its kind, which dynamically suppresses redundant Gaussians during training to enhance generalization; and (2) Edge-Guided Splitting Strategy (ESS), which adaptively refines Gaussians based on image gradients to accurately recover high-frequency geometric details while maintaining low modeling complexity. Both components integrate seamlessly into the standard 3DGS pipeline without requiring additional supervision or pretraining. Extensive experiments on sparse-view benchmarks—Blender, LLFF, and DTU—demonstrate consistent state-of-the-art performance: our method significantly mitigates floating artifacts and substantially improves depth accuracy and surface fidelity.

Technology Category

• Computer Vision: 3D Computer Vision
• Machine Learning: Learning on the Edge & Model Compression
• Knowledge Representation and Reasoning: Geometric, Spatial, and Temporal Reasoning

Application Category

• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking
• Graph Algorithms and Modeling for the Web: Graph embeddings and representation learning for Web-related graphs
• Responsible Web: Data and user privacy-enhancing technologies for the Web

Abstract

Although 3D Gaussian Splatting (3DGS) has demonstrated promising results in novel view synthesis, its performance degrades dramatically with sparse inputs and generates undesirable artifacts. As the number of training views decreases, the novel view synthesis task degrades to a highly under-determined problem such that existing methods suffer from the notorious overfitting issue. Interestingly, we observe that models with fewer Gaussian primitives exhibit less overfitting under sparse inputs. Inspired by this observation, we propose a Random Dropout Regularization (RDR) to exploit the advantages of low-complexity models to alleviate overfitting. In addition, to remedy the lack of high-frequency details for these models, an Edge-guided Splitting Strategy (ESS) is developed. With these two techniques, our method (termed DropoutGS) provides a simple yet effective plug-in approach to improve the generalization performance of existing 3DGS methods. Extensive experiments show that our DropoutGS produces state-of-the-art performance under sparse views on benchmark datasets including Blender, LLFF, and DTU. The project page is at: https://xuyx55.github.io/DropoutGS/.

Problem

• Improves sparse-view rendering in 3D Gaussian Splatting
• Reduces overfitting with Random Dropout Regularization
• Enhances high-frequency details via Edge-guided Splitting

Innovation

• Random Dropout Regularization reduces overfitting
• Edge-guided Splitting Strategy enhances details
• DropoutGS improves 3DGS sparse-view performance