Research map / 3D vision / Gaussian splatting
CCL-LGS: Contrastive Codebook Learning for 3D Language Gaussian Splatting
Lei Tian, Xiaomin Li, Liqian Ma, Hao Yin, Zirui Zheng, Hefei Huang, Taiqing Li, Huchuan Lu, Xu Jia
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 |
|---|---|---|---|
| add pull loss to the codebook baseline (ablation) | codebook baseline · LERF Avg. | mIoU: 56.4 -> 59.5 | Improved |
| add push loss to the codebook baseline (ablation) | codebook baseline · LERF Avg. | mIoU: 56.4 -> 60.4 | Improved |
| add both pull and push losses (full CCL) (ablation) | codebook baseline · LERF Avg. | mIoU: 56.4 -> 65.6 | Improved |
| add push loss on top of pull loss (ablation) | baseline w/ pull loss · LERF Avg. | mIoU: 59.5 -> 65.6 | Improved |
| add pull loss on top of push loss (ablation) | baseline w/ push loss · LERF Avg. | mIoU: 60.4 -> 65.6 | Improved |
| use two merged mask scales instead of one merged single scale (ablation) | CCL-LGS · LERF Avg. | mIoU: 57.4 -> 65.6 | Improved |
| use two merged mask scales instead of one merged single scale (ablation) | CCL-LGS · LERF Ramen | mIoU: 44.3 -> 62.3 | Improved |
| use CCL-LGS instead of 3D VL-GS | 3D VL-GS · LERF Avg. | mIoU: 62.0 -> 65.6 | Improved |
| use CCL-LGS instead of 3D VL-GS | 3D VL-GS · 3D-OVS Avg. | mIoU: 96.9 -> 95.2 | Hurt |
| use CCL-LGS instead of LangSplat | LangSplat · LERF Avg. | mIoU: 51.4 -> 65.6 | Improved |
| use CCL-LGS instead of LangSplat | LangSplat · 3D-OVS Avg. | mIoU: 93.2 -> 95.2 | Improved |
| use CCL-LGS instead of LangSplat | LangSplat · LERF Ramen preprocessing+training time | total minutes: 292min -> 90min | Improved |
Earlier methods this paper compares with
Named baselines from earlier accepted papers in this paper's results tables: better than 1 of 1 on most shared results.
- LangSplat — LangSplat: 3D Language Gaussian Splatting (CVPR 2024) · better on 10 of 11 shared results
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
- Tackling View-Dependent Semantics in 3D Language Gaussian Splatting (ICML 2025 · similarity 0.91)
- Bootstraping Clustering of Gaussians for View-consistent 3D Scene Understanding (AAAI 2025 · similarity 0.89)
- Unifying 2D and 3D Vision-Language Understanding (ICML 2025 · similarity 0.89)
- Chain of Semantics Programming in 3D Gaussian Splatting Representation for 3D Vision Grounding (CVPR 2025 · similarity 0.88)
- 3D Vision-Language Gaussian Splatting (ICLR 2025 · similarity 0.88)
- FastLGS: Speeding Up Language Embedded Gaussians with Feature Grid Mapping (AAAI 2025 · similarity 0.88)
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.
