RoSplat: Robust Feed-Forward Pixel-wise Gaussian Splatting for Varying Input Views and High-Resolution Rendering
For researchers in novel-view synthesis, this work solves two key artifacts in pixel-wise Gaussian splatting, enabling robust high-resolution rendering with varying input views.
RoSplat addresses over-brightness and hole artifacts in feed-forward 3D Gaussian Splatting for novel-view synthesis. It introduces alpha normalization for brightness consistency across varying input views and a 3D sampling regularizer for better Gaussian scale estimation, significantly improving baseline models.
Generalizable 3D Gaussian Splatting has recently emerged as an efficient approach for novel-view synthesis, enabling feed-forward synthesis from only a few input views. However, existing pixel-wise feed-forward methods suffer from over-bright renderings when the number of input views varies during inference, as well as insufficient supervision for accurate Gaussian scale estimation, which leads to hole artifacts, particularly in high-resolution renderings. To address these issues, we identify that the over-brightness is caused by the varying number of overlapping Gaussians and propose a simple alpha normalization strategy to maintain brightness consistency across different number of input views. In addition, we introduce an auxiliary 3D sampling-based regularizer to improve Gaussian scale estimation, thereby mitigating hole artifacts in high-resolution rendering. Experiments on benchmark datasets demonstrate that our method significantly improves baseline models under varying input-view and high-resolution rendering settings.