Research map / 3D vision / 3D generation
Wonder3D++: Cross-Domain Diffusion for High-Fidelity 3D Generation From a Single Image
Yuxiao Yang, Xiaoxiao Long, Zhiyang Dou, Cheng Chung Lin, Yuan Liu, Qingsong Yan, Yuexin Ma, Haoqian Wang, Zhiqiang Wu, Wei Yin
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 cross-domain attention between normal and color domains (ablation) | cross-domain diffusion model · in-the-wild images (qualitative) | normal-color consistency | Improved |
| jointly generate normals and colors instead of sequential normal-to-rgb (ablation) | cross-domain diffusion model · in-the-wild images (qualitative) | color fidelity | Improved |
| jointly generate normals and colors instead of sequential rgb-to-normal (ablation) | cross-domain diffusion model · in-the-wild images (qualitative) | normal geometry | Improved |
| add concave topology check choosing sphere versus Poisson initialization (ablation) | cascaded mesh extraction · complex-topology objects (qualitative) | initial topology | Improved |
| add geometry-aware normal loss (ablation) | mesh reconstruction · in-the-wild images (qualitative) | geometric detail | Improved |
| set camera type switcher to match the input camera (ablation) | Wonder3D++ · orthogonal and perspective inputs (qualitative) | distortion and texture quality | Improved |
| add multi-domain pre-training stage before cross-domain fine-tuning (ablation) | multi-stage training · in-the-wild images (qualitative) | structural consistency and fidelity | Improved |
| add iterative refinement with cross-domain multi-view enhancement module (ablation) | coarse mesh · in-the-wild images (qualitative) | surface detail | Improved |
| replace Wonder3D SDF-based geometry fusion with cascaded mesh extraction (ablation) | reconstruction strategy · in-the-wild images (qualitative) | mesh smoothness and detail | Improved |
| replace Wonder3D with Wonder3D++ (orthogonal setting) | Wonder3D · GSO single-view reconstruction | Chamfer Distance: 0.0222 -> 0.0193 | Improved |
| replace Wonder3D with Wonder3D++ (orthogonal setting) | Wonder3D · GSO single-view reconstruction | Volume IoU: 0.5521 -> 0.6402 | Improved |
| replace Wonder3D with Wonder3D++ (orthogonal setting) | Wonder3D · GSO novel view synthesis | PSNR: 22.92 -> 23.95 | Improved |
Earlier methods this paper compares with
Named baselines from earlier accepted papers in this paper's results tables: better than 5 of 5 on most shared results.
- SyncDreamer — SyncDreamer: Generating Multiview-consistent Images from a Single-view Image (ICLR 2024) · better on 3 of 3 shared results
- Unique3D — Unique3D: High-Quality and Efficient 3D Mesh Generation from a Single Image (NeurIPS 2024) · better on 3 of 3 shared results
- Magic123 — Magic123: One Image to High-Quality 3D Object Generation Using Both 2D and 3D Diffusion Priors (ICLR 2024) · better on 1 of 1 shared result
- Realfusion — RealFusion: 360deg Reconstruction of Any Object From a Single Image (CVPR 2023) · better on 1 of 1 shared result
- One-2-3-45 — One-2-3-45: Any Single Image to 3D Mesh in 45 Seconds without Per-Shape Optimization (NeurIPS 2023) · 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
- Wonder3D: Single Image to 3D using Cross-Domain Diffusion (CVPR 2024 · similarity 0.98)
- One-2-3-45++: Fast Single Image to 3D Objects with Consistent Multi-View Generation and 3D Diffusion (CVPR 2024 · similarity 0.91)
- Unique3D: High-Quality and Efficient 3D Mesh Generation from a Single Image (NeurIPS 2024 · similarity 0.91)
- Cycle3D: High-quality and Consistent Image-to-3D Generation via Generation-Reconstruction Cycle (AAAI 2025 · similarity 0.90)
- DreamCraft3D: Hierarchical 3D Generation with Bootstrapped Diffusion Prior (ICLR 2024 · similarity 0.89)
- Magic123: One Image to High-Quality 3D Object Generation Using Both 2D and 3D Diffusion Priors (ICLR 2024 · similarity 0.89)
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.
