Research map / 3D vision / 3D generation
Towards High-Fidelity 3D Portrait Generation with Rich Details by Cross-View Prior-Aware Diffusion
Haoran Wei, Wencheng Han, Xingping Dong, Jianbing Shen
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 Geo-Block to Explicit-Branch (ablation) | HPDM · HPDM ablation | LPIPS: 0.4013 -> 0.0685 | Improved |
| add Implicit-Branch (ablation) | HPDM · HPDM ablation | LPIPS: 0.0685 -> 0.0405 | Improved |
| add Implicit-Branch (ablation) | HPDM · HPDM ablation | SSIM: 0.8024 -> 0.8732 | Improved |
| replace random noise with MV-NRS (ablation) | SDS optimization · MV-NRS ablation | NIQE: 4.1552 -> 3.9143 | Improved |
| replace random noise with MV-NRS (ablation) | SDS optimization · MV-NRS ablation | Discriminator Score: 0.6104 -> 0.7263 | Improved |
| use only resampled noise without anchor optimization (ablation) | SDS optimization · MV-NRS ablation | NIQE: 4.1552 -> 4.1324 | No clear effect |
| train longer (Detail variant, 3h vs 1h) | Portrait Diffusion · Single-image 3D portraits (30 heads) | CLIP-I: 0.9982 -> 0.9986 | No clear effect |
| add diffusion fine-tuning with Detail Preservation Network to GAN init (ablation) | Geometry restoration · qualitative | geometry quality | Improved |
| use Portrait Diffusion vs ID-Sculpt | Portrait Diffusion · Single-image 3D portraits (30 heads) | ID: 0.3045 -> 0.3438 | Improved |
Earlier methods this paper compares with
Named baselines from earlier accepted papers in this paper's results tables: better than 3 of 3 on most shared results.
- ID-Sculpt — ID-Sculpt: ID-aware 3D Head Generation from Single In-the-wild Portrait Image (AAAI 2025) · better on 3 of 3 shared results
- DreamCraft3D — DreamCraft3D: Hierarchical 3D Generation with Bootstrapped Diffusion Prior (ICLR 2024) · better on 3 of 3 shared results
- Arc2Avatar — Arc2Avatar: Generating Expressive 3D Avatars from a Single Image via ID Guidance (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
- DiffPortrait3D: Controllable Diffusion for Zero-Shot Portrait View Synthesis (CVPR 2024 · similarity 0.90)
- Coherent 3D Scene Diffusion From a Single RGB Image (NeurIPS 2024 · similarity 0.89)
- Cycle3D: High-quality and Consistent Image-to-3D Generation via Generation-Reconstruction Cycle (AAAI 2025 · similarity 0.89)
- Denoising Diffusion via Image-Based Rendering (ICLR 2024 · similarity 0.89)
- Diffusion$^2$: Dynamic 3D Content Generation via Score Composition of Video and Multi-view Diffusion Models (ICLR 2025 · similarity 0.89)
- Era3D: High-Resolution Multiview Diffusion using Efficient Row-wise Attention (NeurIPS 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.
