Lang3D-XL: Language Embedded 3D Gaussians for Large-scale Scenes Article Swipe

MAKE IMPACT

Krakovsky, Shai , Fiebelman, Gal , Benaim, Sagie , Averbuch-Elor, Hadar · YOU? · · 2025 · Open Access ·

Embedding a language field in a 3D representation enables richer semantic understanding of spatial environments by linking geometry with descriptive meaning. This allows for a more intuitive human-computer interaction, enabling querying or editing scenes using natural language, and could potentially improve tasks like scene retrieval, navigation, and multimodal reasoning. While such capabilities could be transformative, in particular for large-scale scenes, we find that recent feature distillation approaches cannot effectively learn over massive Internet data due to challenges in semantic feature misalignment and inefficiency in memory and runtime. To this end, we propose a novel approach to address these challenges. First, we introduce extremely low-dimensional semantic bottleneck features as part of the underlying 3D Gaussian representation. These are processed by rendering and passing them through a multi-resolution, feature-based, hash encoder. This significantly improves efficiency both in runtime and GPU memory. Second, we introduce an Attenuated Downsampler module and propose several regularizations addressing the semantic misalignment of ground truth 2D features. We evaluate our method on the in-the-wild HolyScenes dataset and demonstrate that it surpasses existing approaches in both performance and efficiency.

Type

article

Landing Page

http://arxiv.org/abs/2512.07807

PDF

https://arxiv.org/pdf/2512.07807

OA Status

green

OpenAlex ID

https://openalex.org/W7113916048

Raw OpenAlex JSON

OpenAlex ID

https://openalex.org/W7113916048

Canonical identifier for this work in OpenAlex

Title

Lang3D-XL: Language Embedded 3D Gaussians for Large-scale Scenes

Work title

Type

article

OpenAlex work type

Publication year

2025

Year of publication

Publication date

2025-12-08

Full publication date if available

Authors

Krakovsky, Shai, Fiebelman, Gal, Benaim, Sagie, Averbuch-Elor, Hadar

List of authors in order

Landing page

https://arxiv.org/abs/2512.07807

Publisher landing page

PDF URL

https://arxiv.org/pdf/2512.07807

Direct link to full text PDF

Open access

Yes

Whether a free full text is available

OA status

green

Open access status per OpenAlex

OA URL

https://arxiv.org/pdf/2512.07807

Direct OA link when available

Concepts

Computer science, Artificial intelligence, Embedding, Bottleneck, Rendering (computer graphics), Ground truth, Heuristics, Feature (linguistics), Natural language, Representation (politics), Semantic feature, Natural language processing, Natural language understanding, Field (mathematics), Semantic gap, Semantic computing, Hash function, Bridging (networking), Feature extraction, Computer vision, Semantic interpretation, Search engine indexing, Exploit, Visual reasoning, Machine learning, Semantics (computer science), Semantic mapping, Annotation, Robustness (evolution), Visualization, Image editing, Hash table, Semantic search, Word embedding, Gaussian

Top concepts (fields/topics) attached by OpenAlex

Cited by

0

Total citation count in OpenAlex

Full payload