mlsanityData to 14 Oct 2026

Tackling View-Dependent Semantics in 3D Language Gaussian Splatting

arXiv 2505.24746 ↗First posted 30 May 2025cs.CVcited by 16

Not on the map. Why?

Abstract (arXiv)

Recent advancements in 3D Gaussian Splatting (3D-GS) enable high-quality 3D scene reconstruction from RGB images. Many studies extend this paradigm for language-driven open-vocabulary scene understanding. However, most of them simply project 2D semantic features onto 3D Gaussians and overlook a fundamental gap between 2D and 3D understanding: a 3D object may exhibit various semantics from different viewpoints--a phenomenon we term view-dependent semantics. To address this challenge, we propose LaGa (Language Gaussians), which establishes cross-view semantic connections by decomposing the 3D scene into objects. Then, it constructs view-aggregated semantic representations by clustering semantic descriptors and reweighting them based on multi-view semantics. Extensive experiments demonstrate that LaGa effectively captures key information from view-dependent semantics, enabling a more comprehensive understanding of 3D scenes. Notably, under the same settings, LaGa achieves a significant improvement of +18.7% mIoU over the previous SOTA on the LERF-OVS dataset. Our code is available at: https://github.com/SJTU-DeepVisionLab/LaGa.

Key facts. Cited by 16 papers in the map (data through 14 Oct 2026). Builds on 57 of its 61 references that are in the map. First posted on arXiv 30 May 2025.

Builds on

57 of 61 references.

References · 57

  1. 2021Learning Transferable Visual Models From Natural Language Supervisioncited by 33,744
  2. 2023Segment Anythingcited by 8,234
  3. 2021Emerging Properties in Self-Supervised Vision Transformerscited by 6,968
  4. 20233D Gaussian Splatting for Real-Time Radiance Field Renderingcited by 6,226
  5. 2022Instant Neural Graphics Primitives with a Multiresolution Hash Encodingcited by 3,723
  6. 2017ScanNet: Richly-annotated 3D Reconstructions of Indoor Scenescited by 3,117
  7. 2021Mip-NeRF 360: Unbounded Anti-Aliased Neural Radiance Fieldscited by 2,032
  8. 2021Mip-NeRF: A Multiscale Representation for Anti-Aliasing Neural Radiance Fieldscited by 1,964
  9. 2021Plenoxels: Radiance Fields without Neural Networkscited by 1,613
  10. 2022TensoRF: Tensorial Radiance Fieldscited by 1,280
  11. 2020NeRF in the Wild: Neural Radiance Fields for Unconstrained Photo Collectionscited by 1,232
  12. 20234D Gaussian Splatting for Real-Time Dynamic Scene Renderingcited by 1,157
38 more
  1. 2021Direct Voxel Grid Optimization: Super-fast Convergence for Radiance Fields Reconstructioncited by 964
  2. 2020NeRF++: Analyzing and Improving Neural Radiance Fieldscited by 861
  3. 2023Deformable 3D Gaussians for High-Fidelity Monocular Dynamic Scene Reconstructioncited by 759
  4. 2024LGM: Large Multi-View Gaussian Model for High-Resolution 3D Content Creationcited by 739
  5. 2021KiloNeRF: Speeding up Neural Radiance Fields with Thousands of Tiny MLPscited by 673
  6. 2023Zip-NeRF: Anti-Aliased Grid-Based Neural Radiance Fieldscited by 671
  7. 2023LERF: Language Embedded Radiance Fieldscited by 543
  8. 2021Baking Neural Radiance Fields for Real-Time View Synthesiscited by 468
  9. 2021In-Place Scene Labelling and Understanding with Implicit Scene Representationcited by 415
  10. 2023LangSplat: 3D Language Gaussian Splattingcited by 415
  11. 2022Decomposing NeRF for Editing via Feature Field Distillationcited by 371
  12. 2023Feature 3DGS: Supercharging 3D Gaussian Splatting to Enable Distilled Feature Fieldscited by 364
  13. 2023GaussianEditor: Swift and Controllable 3D Editing with Gaussian Splattingcited by 335
  14. 2023GaussianDreamer: Fast Generation from Text to 3D Gaussians by Bridging 2D and 3D Diffusion Modelscited by 255
  15. 2021DONeRF: Towards Real-Time Rendering of Compact Neural Radiance Fields using Depth Oracle Networkscited by 232
  16. 2021NeX: Real-time View Synthesis with Neural Basis Expansioncited by 216
  17. 2020AutoInt: Automatic Integration for Fast Neural Volume Renderingcited by 204
  18. 2022Neural Feature Fusion Fields: 3D Distillation of Self-Supervised 2D Image Representationscited by 196
  19. 20244D-Rotor Gaussian Splatting: Towards Efficient Novel View Synthesis for Dynamic Scenescited by 189
  20. 2022Panoptic Lifting for 3D Scene Understanding with Neural Fieldscited by 174
  21. 2023GaussianEditor: Editing 3D Gaussians Delicately with Text Instructionscited by 168
  22. 2023Segment Any 3D Gaussianscited by 161
  23. 2022Panoptic NeRF: 3D-to-2D Label Transfer for Panoptic Urban Scene Segmentationcited by 152
  24. 2024OpenGaussian: Towards Point-Level 3D Gaussian-based Open Vocabulary Understandingcited by 145
  25. 2023Weakly Supervised 3D Open-vocabulary Segmentationcited by 135
  26. 2021NeSF: Neural Semantic Fields for Generalizable Semantic Segmentation of 3D Scenescited by 119
  27. 2021Decomposing 3D Scenes into Objects via Unsupervised Volume Segmentationcited by 88
  28. 2023OmniSeg3D: Omniversal 3D Segmentation via Hierarchical Contrastive Learningcited by 75
  29. 2022NeRF-SOS: Any-View Self-supervised Object Segmentation on Complex Scenescited by 71
  30. 2022DM-NeRF: 3D Scene Geometry Decomposition and Manipulation from 2D Imagescited by 69
  31. 2023Contrastive Lift: 3D Object Instance Segmentation by Slow-Fast Contrastive Fusioncited by 64
  32. 2023Segment Anything in 3D with Radiance Fieldscited by 62
  33. 2022Neural Volumetric Object Selectioncited by 61
  34. 2022Interactive Segmentation of Radiance Fieldscited by 60
  35. 2024GaussianObject: High-Quality 3D Object Reconstruction from Four Views with Gaussian Splattingcited by 59
  36. 2024GOI: Find 3D Gaussians of Interest with an Optimizable Open-vocabulary Semantic-space Hyperplanecited by 53
  37. 2024SplatFields: Neural Gaussian Splats for Sparse 3D and 4D Reconstructioncited by 52
  38. 2022NeRF-RPN: A general framework for object detection in NeRFscited by 46

50 most cited of 57 All 57 (JSON)

Built on by

Cited by year (2026 to 14 Oct)

Citing papers per year: 2025: 3; 2026: 13.

16 papers.

Sort:

By area:

  1. 2025Splat Feature Solvercited by 7
  2. 2025ExtrinSplat: Decoupling Geometry and Semantics for Open-Vocabulary Understanding in 3D Gaussian Splattingcited by 6
  3. 2026ReLaGS: Relational Language Gaussian Splattingcited by 6
  4. 2025Vision-Language Memory for Spatial Reasoningcited by 5
  5. 2026EnerGS: Energy-Based Gaussian Splatting with Partial Geometric Priorscited by 2
  6. 2026LESV: Language Embedded Sparse Voxel Fusion for Open-Vocabulary 3D Scene Understandingcited by 1
  7. 2026Open-Vocabulary and Referring Segmentation for 3D Gaussians Using 2D Detectorscited by 1
  8. 2026Relation-Centric Open-Vocabulary 3D Gaussian Segmentationcited by 1
  9. 2026A3R: Agentic Affordance Reasoning via Cross-Dimensional Evidence in 3D Gaussian Scenescited by 0
  10. 2026MLP Splatting: Object-Centric Neural Fieldscited by 0
  11. 2026LEGO: Leveled Language Gaussian Splattingcited by 0
  12. 2026NormLift: From Lifted Features To Semantic Reliability In 3D Gaussian Splattingcited by 0
4 more
  1. 2026CoRef-GS: Cooperative Referring Gaussian Splatting for Multi-Agent Scene Understandingcited by 0
  2. 2026EviSplat: Preserving Multi-View Evidence in 3D Gaussian Splatting for Open-Vocabulary Segmentationcited by 0
  3. 2026RelationVGGT: Visual Geometry Transformers for 3D Spatial Relation Segmentationcited by 0
  4. 2026DIGS: Distribution-Informed Gaussian Splatting for Training-Free Open-Vocabulary 3D Segmentationcited by 0