RobustSplat: Decoupling Densification and Dynamics for Transient-Free 3DGS
This addresses a specific issue in 3D modeling and novel-view synthesis for scenes with moving objects, representing an incremental improvement over prior 3DGS methods.
The paper tackles the problem of artifacts in 3D Gaussian Splatting (3DGS) caused by transient objects, proposing RobustSplat with a delayed Gaussian growth strategy and scale-cascaded mask bootstrapping, which outperforms existing methods on multiple datasets.
3D Gaussian Splatting (3DGS) has gained significant attention for its real-time, photo-realistic rendering in novel-view synthesis and 3D modeling. However, existing methods struggle with accurately modeling scenes affected by transient objects, leading to artifacts in the rendered images. We identify that the Gaussian densification process, while enhancing scene detail capture, unintentionally contributes to these artifacts by growing additional Gaussians that model transient disturbances. To address this, we propose RobustSplat, a robust solution based on two critical designs. First, we introduce a delayed Gaussian growth strategy that prioritizes optimizing static scene structure before allowing Gaussian splitting/cloning, mitigating overfitting to transient objects in early optimization. Second, we design a scale-cascaded mask bootstrapping approach that first leverages lower-resolution feature similarity supervision for reliable initial transient mask estimation, taking advantage of its stronger semantic consistency and robustness to noise, and then progresses to high-resolution supervision to achieve more precise mask prediction. Extensive experiments on multiple challenging datasets show that our method outperforms existing methods, clearly demonstrating the robustness and effectiveness of our method. Our project page is https://fcyycf.github.io/RobustSplat/.