SparseSplat: Towards Applicable Feed-Forward 3D Gaussian Splatting with Pixel-Unaligned Prediction
This addresses the integration of feed-forward 3DGS into downstream reconstruction tasks by reducing redundancy, offering a domain-specific improvement for 3D rendering and reconstruction.
The paper tackled the problem of redundant 3D Gaussian Splatting maps in feed-forward models by proposing SparseSplat, which adaptively adjusts Gaussian density based on scene structure, achieving state-of-the-art rendering quality with only 22% of the Gaussians and maintaining reasonable quality with 1.5%.
Recent progress in feed-forward 3D Gaussian Splatting (3DGS) has notably improved rendering quality. However, the spatially uniform and highly redundant 3DGS map generated by previous feed-forward 3DGS methods limits their integration into downstream reconstruction tasks. We propose SparseSplat, the first feed-forward 3DGS model that adaptively adjusts Gaussian density according to scene structure and information richness of local regions, yielding highly compact 3DGS maps. To achieve this, we propose entropy-based probabilistic sampling, generating large, sparse Gaussians in textureless areas and assigning small, dense Gaussians to regions with rich information. Additionally, we designed a specialized point cloud network that efficiently encodes local context and decodes it into 3DGS attributes, addressing the receptive field mismatch between the general 3DGS optimization pipeline and feed-forward models. Extensive experimental results demonstrate that SparseSplat can achieve state-of-the-art rendering quality with only 22% of the Gaussians and maintain reasonable rendering quality with only 1.5% of the Gaussians. Project page: https://victkk.github.io/SparseSplat-page/.