aiqbench analysis
ExtrinSplat addresses critical inefficiencies in 3D Gaussian Splatting by decoupling geometry from semantics, significantly reducing storage overhead and scene adaptation time while improving open-vocabulary understanding. By leveraging Vision-Language Models to interpret clustered object groups rather than embedding features directly, this approach offers a scalable solution for complex 3D scene analysis tasks.
ExtrinSplat: Decoupling Geometry and Semantics for Open-Vocabulary Understanding in 3D Gaussian Splatting
arXiv:2509.22225v3 Announce Type: replace-cross Abstract: Lifting 2D open-vocabulary understanding into 3D Gaussian Splatting (3DGS) scenes is a critical challenge. Mainstream methods, built on an embedding paradigm, suffer from three key flaws: (i) geometry-semantic inconsistency, where points, rather than objects, serve as the semantic basis, limiting semantic fidelity; (ii) semantic bloat from injecting gigabytes of feature data into the geometry; and (iii) semantic rigidity, as one feature per Gaussian struggles to capture rich polysemy. To overcome these limitations, we introduce ExtrinSplat, a framework built on the extrinsic paradigm that decouples geometry from semantics. Instead of embedding features, ExtrinSplat clusters Gaussians into multi-granularity, overlapping 3D object groups. A Vision-Language Model (VLM) then interprets these groups to generate lightweight textual hypotheses, creating an extrinsic index layer that natively supports complex polysemy. By replacing costly feature embedding with lightweight indices, ExtrinSplat reduces scene adaptation time from hours to minutes and lowers storage overhead by several orders of magnitude. On benchmark tasks for open-vocabulary 3D object selection and semantic segmentation, ExtrinSplat outperforms established embedding-based frameworks, validating the efficacy and efficiency of the proposed extrinsic paradigm.
Abstract:Lifting 2D open-vocabulary understanding into 3D Gaussian Splatting (3DGS) scenes is a critical challenge. Mainstream methods, built on an embedding paradigm, suffer from three key flaws: (i) geometry-semantic inconsistency, where points, rather than objects, serve as the semantic basis, limiting semantic fidelity; (ii) semantic bloat from injecting gigabytes of feature data into the geometry; and (iii) semantic rigidity, as one feature per Gaussian struggles to capture rich polysemy. To overcome these limitations, we introduce ExtrinSplat, a framework built on the extrinsic paradigm that decouples geometry from semantics. Instead of embedding features, ExtrinSplat clusters Gaussians into multi-granularity, overlapping 3D object groups. A Vision-Language Model (VLM) then interprets these groups to generate lightweight textual hypotheses, creating an extrinsic index layer that natively supports complex polysemy. By replacing costly feature embedding with lightweight indices, ExtrinSplat reduces scene adaptation time from hours to minutes and lowers storage overhead by several orders of magnitude. On benchmark tasks for open-vocabulary 3D object selection and semantic segmentation, ExtrinSplat outperforms established embedding-based frameworks, validating the efficacy and efficiency of the proposed extrinsic paradigm.
Computer Vision and Pattern Recognition (cs.CV); Artificial Intelligence (cs.AI)
arXiv:2509.22225 [cs.CV]
(or arXiv:2509.22225v3 [cs.CV] for this version)
https://doi.org/10.48550/arXiv.2509.22225 Focus to learn more arXiv-issued DOI via DataCite
arXivLabs is a framework that allows collaborators to develop and share new arXiv features directly on our website.
Both individuals and organizations that work with arXivLabs have embraced and accepted our values of openness, community, excellence, and user data privacy. arXiv is committed to these values and only works with partners that adhere to them.
Have an idea for a project that will add value for arXiv's community? Learn more about arXivLabs.
Full text under the publisher's open licence · 273 words
Related coverage
OpenSplatGraph: From Dense Semantic Maps to Structured Scene Graphs for Open-Vocabulary Robot Perception
arXiv · Oct 7, 2026
DGSG-Mind: Dynamic 3D Gaussian Scene Graphs for Long-Term Scene Understanding and Grounding
arXiv · Sep 16, 2026
ActiveLang: Active Open-Vocabulary 3D Mapping with Semantic-Uncertainty-Guided Exploration
arXiv · Oct 8, 2026
SCOUT: Semantic scene COverage via Uncertainty-guided Traversal
arXiv · Jun 8, 2026