Search papers, labs, and topics across Lattice.
This paper introduces LocusGS, a novel approach that enhances query-based feed-forward 3D Gaussian Splatting by incorporating spatially grounded tokens, specifically 3D anchor states for each Gaussian query. By refining these anchor states across decoder layers, the method improves spatial coherence and alignment with scene structures, addressing the limitations of existing Gaussian queries that often scatter across distant regions. Experimental results demonstrate that LocusGS significantly enhances rendering quality in novel view synthesis tasks while maintaining the same Gaussian budget, showcasing the effectiveness of explicit anchor states in organizing Gaussian distributions.
Spatially grounded tokens in LocusGS lead to coherent Gaussian distributions that dramatically improve rendering quality in 3D scene synthesis.
Recent query-based feed-forward 3DGS methods represent a scene using learnable queries, each aggregating multi-view evidence and decoding a group of Gaussians. Ideally, different queries should specialize in coherent local regions of the scene. However, we observe that Gaussians decoded from the same query often scatter across distant scene regions, resulting in weak query-level spatial coherence and poor alignment with the scene structure. We attribute this behavior to the purely latent representation of existing Gaussian queries. To address this limitation, we introduce LocusGS, which augments each Gaussian query with a 3D anchor state consisting of a center and a support radius. The anchor state is progressively refined across decoder layers and is used throughout query interaction, multi-view feature aggregation, and Gaussian generation. Specifically, an anchor-to-ray geometric bias guides each query toward spatially relevant image observations, while anchor-centered decoding organizes its Gaussians within a local region. Experiments on novel view synthesis benchmarks show that LocusGS improves rendering quality over query-based Gaussian token baselines under the same Gaussian budget. Further analysis shows that the learned anchors form coherent spatial layouts and lead to more structured Gaussian distributions, demonstrating that explicit anchor states improve the spatial organization. Our project page: https://leo-frank.github.io/LocusGS_viewer.
