F4Splat: Feed-Forward Predictive Densification for Feed-Forward 3D Gaussian Splatting

TL;DR

F4Splat introduces a predictive densification approach for 3D Gaussian Splatting that adaptively allocates Gaussians based on spatial complexity, reducing redundancy and improving rendering quality in real-time 3D reconstruction.

Contribution

The paper proposes a novel densification-score-guided strategy for adaptive Gaussian allocation in feed-forward 3D splatting, enabling explicit control over Gaussian count and reducing redundancy.

Findings

Achieves higher quality novel-view synthesis with fewer Gaussians.

Reduces Gaussian redundancy across views and in simple regions.

Maintains real-time performance with improved reconstruction fidelity.

Abstract

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.