Research map / 3D vision / 3D generation
ROAR-3D: Routing Arbitrary Views for High-Fidelity 3D Generation
Hanxiao Sun, Mingxin Yang, Shuhui Yang, Zebin He, Xintong Han, Hongbo Fu, Chunchao Guo, Wenhan Luo
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 token-wise view router instead of concatenating view tokens (ablation) | ROAR-3D stage 1 · Anyview-200 | CD: 39.412 -> 28.143 | Improved |
| add token-wise view router (ablation) | ROAR-3D stage 1 · Anyview-200 | F1(0.05): 66.3 -> 73.5 | Improved |
| add dual-stream attention (ablation) | ROAR-3D stage 1 · Anyview-200 | CD: 28.143 -> 26.732 | Improved |
| add dual-stream attention (ablation) | ROAR-3D stage 1 · Anyview-200 | Uni-I: 0.335 -> 0.339 | Improved |
| add orientation perturbation (ablation) | ROAR-3D stage 1 · Anyview-200 | CD: 26.732 -> 25.372 | Improved |
| add orientation perturbation (ablation) | ROAR-3D stage 1 · Anyview-200 | F1(0.05): 75.1 -> 76.8 | Improved |
| add stage 2 geometric refinement | ROAR-3D · Held-out test set | CD: 25.372 -> 21.039 | Improved |
| add stage 2 geometric refinement | ROAR-3D · Held-out test set | F1(0.05): 76.8 -> 81.6 | Improved |
| use ROAR-3D instead of strongest baseline ReconViaGen | ROAR-3D stage 2 · Held-out test set | CD: 32.055 -> 21.039 (34% reduction) | Improved |
| condition on multiple views instead of single view | Hunyuan3D 2.0 (MV variant) · Held-out test set | CD: 35.421 -> 33.954 | No clear effect |
| increase test-time input views beyond training range (1-4 to 8) | ROAR-3D · Anyview-200 subset | qualitative geometry | 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.
- ReconViaGen — ReconViaGen: Towards Accurate Multi-view 3D Object Reconstruction via Generation (ICLR 2026) · better on 5 of 5 shared results
- VGGT — VGGT: Visual Geometry Grounded Transformer (CVPR 2025) · 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
- One View Is Enough: In-the-Wild Monocular Pretraining for Novel View Generation (NeurIPS 2026 · similarity 0.88)
- Sculpt3D: Multi-View Consistent Text-to-3D Generation with Sparse 3D Prior (CVPR 2024 · similarity 0.87)
- Cue3D: Quantifying the Role of Image Cues in Single-Image 3D Generation (NeurIPS 2025 · similarity 0.87)
- NAMVIS: Next-Scale Autoregressive Multi-View Image Synthesis (NeurIPS 2026 · similarity 0.87)
- MVD-Fusion: Single-view 3D via Depth-consistent Multi-view Generation (CVPR 2024 · similarity 0.87)
- MVGenMaster: Scaling Multi-View Generation from Any Image via 3D Priors Enhanced Diffusion Model (CVPR 2025 · similarity 0.87)
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.
