Research map / 3D vision / Gaussian splatting
LERF: Language Embedded Radiance Fields
Justin Kerr, Chung Min Kim, Ken Goldberg, Angjoo Kanazawa, Matthew Tancik
Read the paperCheck your idea against 137,217 papers
AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.
Measured effects reported in the paper
| What was changed | Where | Measured | Effect |
|---|---|---|---|
| Use LERF's 3D CLIP field instead of distilling LSeg features into 3D | LERF · Localization, 5 scenes | localization accuracy: 18.0% -> 80.3% | Improved |
| Use LERF instead of 2D open-vocab detector OWL-ViT on rendered views | LERF · Localization, 5 scenes | localization accuracy: 54.8% -> 80.3% overall; ramen 92.5% -> 62.5% | Mixed |
| Add DINO regularization to the language field (ablation) | LERF · relevancy map quality (qualitative) | map smoothness and boundaries | Improved |
| Use multi-scale CLIP supervision instead of single fixed 15% scale (ablation) | LERF · relevancy queries at all scales (qualitative) | query handling | Improved |
| Use CLIP-based LERF instead of LSeg in 3D for existence determination on long-tail labels | LERF · Existence determination (5 scenes) | precision-recall | Improved |
| Use more specific query wording (e.g. blue dish soap vs dish soap) | LERF queries · qualitative relevancy maps | relevancy localization | Improved |
Later papers that compared with this method
19 later accepted papers report LERF as a baseline in their results tables; 18 of them report a better number on most of the results they share with it.
- LangSurf: Language-Embedded Surface Gaussians for 3D Scene Understanding (TPAMI 2026 · better on 2 of 2 shared results)
- ReLaGS: Relational Language Gaussian Splatting (CVPR 2026 · better on 1 of 1 shared result)
- Language Embedded 3D Gaussians for Open-Vocabulary Scene Querying (TPAMI 2025 · better on 5 of 5 shared results)
- Functionality Understanding and Segmentation in 3D Scenes (CVPR 2025 · better on 4 of 4 shared results)
- OpenHype: Hyperbolic Embeddings for Hierarchical Open-Vocabulary Radiance Fields (NeurIPS 2025 · better on 3 of 3 shared results)
- Bootstraping Clustering of Gaussians for View-consistent 3D Scene Understanding (AAAI 2025 · better on 2 of 2 shared results)
- LUDVIG: Learning-Free Uplifting of 2D Visual Features to Gaussian Splatting Scenes (ICCV 2025 · better on 2 of 2 shared results)
- FastLGS: Speeding Up Language Embedded Gaussians with Feature Grid Mapping (AAAI 2025 · better on 0 of 1 shared result)
11 more not listed.
Matched by method name in results tables read by AI. Authors' own tables tend to favour the newer method, and a comparison can be missed when a method is renamed.
Closest papers by meaning
- LangSplat: 3D Language Gaussian Splatting (CVPR 2024 · similarity 0.89)
- Rethinking Open-Vocabulary Segmentation of Radiance Fields in 3D Space (AAAI 2025 · similarity 0.89)
- LLaNA: Large Language and NeRF Assistant (NeurIPS 2024 · similarity 0.88)
- Language Embedded 3D Gaussians for Open-Vocabulary Scene Understanding (CVPR 2024 · similarity 0.87)
- FastLGS: Speeding Up Language Embedded Gaussians with Feature Grid Mapping (AAAI 2025 · similarity 0.86)
- GP-NeRF: Generalized Perception NeRF for Context-Aware 3D Scene Understanding (CVPR 2024 · similarity 0.86)
Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.
