Research map / 3D vision / Gaussian splatting
DASAM3D: A Unified Foundation Model for Enhanced 3D Scene Reconstruction and Segmentation
RONNY XAVIER VELASTEGUI SANDOVAL, Max Pfingsthorn, Sezer Karaoglu, Theo Gevers
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 |
|---|---|---|---|
| fuse SAM 3D layout into DA3 | DA3 · MipNeRF-360 (6 scenes) 3D reconstruction | Chamfer Distance (cm): 10.76 -> 10.10 | Improved |
| fuse SAM 3D layout into DA3 | DA3 · DL3DV-10K NVS | PSNR (dB): 26.64 -> 27.08 | Improved |
| fuse SAM 3D layout into DA3 | DA3 · MipNeRF-360 (6 scenes) NVS | LPIPS: 0.237 -> 0.234 | No clear effect |
| add per-object local alignment after global alignment (ablation) | DASAM3D · MipNeRF-360 Room | Chamfer Distance (cm): 11.41 -> 9.86 | Improved |
| add per-object local alignment after global alignment (ablation) | DASAM3D · MipNeRF-360 Counter | F-score (%): 55.2 -> 63.4 | Improved |
| add per-object local alignment after global alignment (ablation) | DASAM3D · MipNeRF-360 Kitchen | F-score (%): 56.7 -> 64.1 | Improved |
| increase number of SAM 3D objects K from 1 to 8 (ablation) | DASAM3D · MipNeRF-360 Counter | Chamfer Distance (cm): 9.41 -> 7.58 | Improved |
| add quality-based object filtering (ablation) | DASAM3D · MipNeRF-360 Counter | PSNR (dB): 24.78 -> 25.14 | Improved |
| add quality-based object filtering (ablation) | DASAM3D · MipNeRF-360 Counter | N-MAE (deg): 17.4 -> 15.6 | Improved |
| add depth loss in Stage 3 (ablation) | DASAM3D · MipNeRF-360 Counter | Abs Rel: 0.142 -> 0.098 | Improved |
| add normal loss in Stage 3 (ablation) | DASAM3D · MipNeRF-360 Counter | N-MAE (deg): 21.3 -> 15.6 | Improved |
| add RGB consistency loss in Stage 3 (ablation) | DASAM3D · MipNeRF-360 Counter | PSNR (dB): 24.23 -> 25.14 | Improved |
Earlier methods this paper compares with
Named baselines from earlier accepted papers in this paper's results tables: better than 2 of 2 on most shared results.
- VGGT — VGGT: Visual Geometry Grounded Transformer (CVPR 2025) · better on 2 of 2 shared results
- SAM 3D — SAM 3D: 3Dfy Anything in Images (CVPR 2026) · better on 1 of 1 shared result
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
- SAM 3D: 3Dfy Anything in Images (CVPR 2026 · similarity 0.89)
- Fast-SAM3D: 3Dfy Anything in Images but Faster (ICML 2026 · similarity 0.89)
- RealX3D: A Physically-Degraded 3D Benchmark for Multi-view Visual Restoration and Reconstruction (IJCV 2026 · similarity 0.87)
- Photo3D: Advancing Photorealistic 3D Generation through Structure-Aligned Detail Enhancement (CVPR 2026 · similarity 0.87)
- ARM: Appearance Reconstruction Model for Relightable 3D Generation (CVPR 2025 · similarity 0.86)
- EA3D: Online Open-World 3D Object Extraction from Streaming Videos (NeurIPS 2025 · 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.
