F4Splat: Feed-Forward Predictive Densification for Feed-Forward 3D Gaussian Splatting
This work addresses inefficiencies in 3D reconstruction for real-time rendering applications, offering an incremental improvement in Gaussian allocation and control.
The paper tackles the problem of redundant and uniformly allocated Gaussians in feed-forward 3D Gaussian splatting methods, which lack control over Gaussian numbers while maintaining fidelity. The result is F4Splat, which uses a densification-score-guided allocation strategy to reduce Gaussians significantly while achieving superior novel-view synthesis performance.
Feed-forward 3D Gaussian Splatting methods enable single-pass reconstruction and real-time rendering. However, they typically adopt rigid pixel-to-Gaussian or voxel-to-Gaussian pipelines that uniformly allocate Gaussians, leading to redundant Gaussians across views. Moreover, they lack an effective mechanism to control the total number of Gaussians while maintaining reconstruction fidelity. To address these limitations, we present F4Splat, which performs Feed-Forward predictive densification for Feed-Forward 3D Gaussian Splatting, introducing a densification-score-guided allocation strategy that adaptively distributes Gaussians according to spatial complexity and multi-view overlap. Our model predicts per-region densification scores to estimate the required Gaussian density and allows explicit control over the final Gaussian budget without retraining. This spatially adaptive allocation reduces redundancy in simple regions and minimizes duplicate Gaussians across overlapping views, producing compact yet high-quality 3D representations. Extensive experiments demonstrate that our model achieves superior novel-view synthesis performance compared to prior uncalibrated feed-forward methods, while using significantly fewer Gaussians.