Research map / 3D vision / Gaussian splatting
MLLMSplat: A 2D MLLM-Powered Framework for 3D Gaussian Splatting Understanding, Generation, and Editing
Jingqiao Xiu, Can Wang, Dong Xu
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 Gaussian tokenizer instead of built-in visual tokenizer | Qwen2.5-VL-3B · ScanQA val | CIDEr: 55.85 -> 62.69 | Improved |
| use Gaussian tokenizer instead of built-in visual tokenizer | LLaVA-Video-7B · ScanQA val | CIDEr: 87.05 -> 93.71 | Improved |
| use Gaussian tokenizer instead of built-in visual tokenizer | LLaVA-Video-7B · SQA3D test | EM: 50.01 -> 53.49 | Improved |
| add dual rotary positional encoding (GaPE+RoPE) instead of Plucker concat (ablation) | MLLMSplat generation · RealEstate10K | FID: 67.34 -> 50.07 | Improved |
| add dual-stream 3DGS decoder (ablation) | MLLMSplat generation · RealEstate10K | FID: 53.69 -> 50.07 | Improved |
| add velocity refinement during sampling (ablation) | MLLMSplat generation · RealEstate10K | FID: 58.21 -> 50.07 | Improved |
| add DRoPE (ablation) | MLLMSplat generation · DL3DV-10K | CLIP score: 23.20 -> 27.68 | Improved |
| MLLMSplat vs best baseline Prometheus | MLLMSplat · RealEstate10K | FID: 63.40 -> 50.07 | Improved |
| MLLMSplat feed-forward editing vs DGE | MLLMSplat · 3DGS editing (10 scenes) | CLIP-Dir: 15.38 -> 22.40 | Improved |
| MLLMSplat feed-forward editing vs DGE | MLLMSplat · 3DGS editing (10 scenes) | Avg-Time: 100s -> 25s | 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.
- Director3D — Director3D: Real-world Camera Trajectory and 3D Scene Generation from Text (NeurIPS 2024) · better on 4 of 4 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
- EditSplat: Multi-View Fusion and Attention-Guided Optimization for View-Consistent 3D Scene Editing with 3D Gaussian Splatting (CVPR 2025 · similarity 0.92)
- SceneSplat++: A Large Dataset and Comprehensive Benchmark for Language Gaussian Splatting (NeurIPS 2025 · similarity 0.92)
- SplatFlow: Multi-View Rectified Flow Model for 3D Gaussian Splatting Synthesis (CVPR 2025 · similarity 0.92)
- GaussianSpa: An "Optimizing-Sparsifying" Simplification Framework for Compact and High-Quality 3D Gaussian Splatting (CVPR 2025 · similarity 0.91)
- 3DitScene: Editing Any Scene via Language-guided Disentangled Gaussian Splatting (ICLR 2025 · similarity 0.90)
- Dr. Splat: Directly Referring 3D Gaussian Splatting via Direct Language Embedding Registration (CVPR 2025 · similarity 0.90)
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.
