Research map / 3D vision / Face and shape
Shapefusion: 3D localized human diffusion models
Rolandos Alexandros Potamias, Michael Tarasiou, Stylianos Ploumpis, Stefanos Zafeiriou
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 |
|---|---|---|---|
| train a diffusion model with random region masking (inpainting-style) instead of latent-space segmentation (SD, LED) | ShapeFusion · MimicMe | FID (mm): 3.39 (LED) -> 0.30 | Improved |
| diffusion model instead of masked VAE (M-VAE) with the same masking technique (ablation) | ShapeFusion · STAR | DIV (mm): 0.72 -> 0.98 | Improved |
| diffusion model instead of M-VAE (ablation) | ShapeFusion · UHM | FID (mm): 1.17 -> 0.53 | Improved |
| use mask-conditioned direct point manipulation instead of optimization-based fitting | ShapeFusion · point manipulation runtime | time (sec): ~22sec -> ~3.2sec | Improved |
| use ShapeFusion for expression editing vs NFR | ShapeFusion · expression editing runtime | time: 60s -> 3s | Improved |
| add regularization to NFR to protect non-edited regions | NFR · expression editing | manipulation quality | Mixed |
Closest papers by meaning
- Human-3Diffusion: Realistic Avatar Creation via Explicit 3D Consistent Diffusion Models (NeurIPS 2024 · similarity 0.88)
- Diffusion Handles Enabling 3D Edits for Diffusion Models by Lifting Activations to 3D (CVPR 2024 · similarity 0.87)
- Denoising Functional Maps: Diffusion Models for Shape Correspondence (CVPR 2025 · similarity 0.87)
- Diffusion 3D Features (Diff3F): Decorating Untextured Shapes with Distilled Semantic Features (CVPR 2024 · similarity 0.87)
- Flow-Corrected Shape Optimization: Taming Manifold Drift in High-Dimensional 3D Models (NeurIPS 2026 · similarity 0.87)
- Morphable Diffusion: 3D-Consistent Diffusion for Single-image Avatar Creation (CVPR 2024 · 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.
