RelateGS is presented, a structured language-aware Gaussian representation that selectively refines semantic and relational evidence for 3D visual grounding and suppresses unreliable observations during semantic feature uplifting and consolidates multi-view object observations into a persistent 3D graph.
Abstract
Language-aware 3D Gaussian Splatting (3DGS) has advanced language-driven 3D scene understanding by incorporating semantic and relational cues into reconstructed scenes. However, relation-aware grounding depends not only on representing such information, but also on whether the underlying scene-side evidence is reliable and sufficient. Multi-view semantic aggregation can incorporate weak observations, while object relations can remain under-observed along the available camera trajectory, leading to ambiguous evidence for target disambiguation. We present RelateGS, a structured language-aware Gaussian representation that selectively refines semantic and relational evidence for 3D visual grounding. RelateGS suppresses unreliable observations during semantic feature uplifting and consolidates multi-view object observations into a persistent 3D graph. The graph further identifies under-observed object relations and guides targeted re-querying of informative existing frames to complement their evidence. Experiments on Sr3D+ and Nr3D demonstrate competitive 3D grounding performance, improving the previous best by 7.9 and 1.7 percentage points at Acc@0.25, respectively.
Learning semantic representations for 3D Gaussian Splatting has recently emerged as a promising direction for 3D scene understanding. However, existing 2D-to-3D semantic lifting approaches rely primarily on 2D segmentation supervision and lack object-level priors, causing the learned representation to be driven by loca...
Yuntae Jeon, Y. Jeon, Su-Jin Jin et al.· 0 citations
Online semantic 3D Gaussian Splatting (3DGS) enables open-vocabulary scene understanding from streaming observations. However, existing methods recover geometric context missing from image-derived semantic features using an additionally learned 3D network, incurring extra training and runtime overhead. We observe that...
QVR3D is introduced, a query-guided visual reasoning framework that operates on RGB-D observations using off-the-shelf foundation models that integrates instance confidence, attribute consistency, and relationship consistency and supports both single- and multi-object grounding without task-specific adaptation.
Jung-Hao Kao, Jing Huang, Jia-Wei Liao et al.· 0 citations
Understanding and reasoning over 3D scenes from point clouds is essential for a broad range of AI applications. Recent 3D large language models have shown promising results, yet they still face two fundamental limitations: (i) insufficient capability for grounded multi-step reasoning in complex 3D scenes, and (ii) a te...
Yongqi Shan, Yi-Fan Wang, Yun-Zhi Zhuge et al.· Proceedings of the 34th ACM...· 0 citations
EviSplat is introduced, which preserves individual observation features as evidence for later text queries within class-agnostic 3D instances that represent objects, object parts, or background regions and learns a distribution describing which visual appearances its observations support.
Sung-Ho Moon, Kota Shimomura, Junwoo Park et al.· arXiv.org· 0 citations
3D scene graphs support language-guided navigation by organizing physical entities and their containment and spatial relations into hierarchical maps. Building such structured maps requires a scene representation that consistently distinguishes semantic regions and object instances across viewpoints. To this end, we pr...
Jing Gao, Si-Yu Wang, Dong Feng et al.· 0 citations
We use cookies to run the site and, with your consent, for analytics and to show ads.
See our Cookie Policy.