Research map / 3D vision / NeRF and radiance fields
HumanGen: Generating Human Radiance Fields With Explicit Priors
Read the paperCheck your idea against 108,967 papers
AI-extracted summary. Written by AI from the paper's title only (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 PIFuHD reconstruction features to the geometry branch (ablation) | HumanGen · Geometry generation | Depth Diff: 0.0383 -> 0.0145 | Helped |
| Use SDF instead of density output (ablation) | HumanGen geometry generation · Geometry generation | Depth Diff: 0.0183 -> 0.0145 | Helped |
| Add frontal and back consistency supervision (ablation) | HumanGen texture generation · Texture generation | FID: 34.52 -> 20.97 | Helped |
| Add frontal consistency supervision (ablation) | HumanGen texture generation · Texture generation | FID: 25.75 -> 20.97 | Helped |
| Add back consistency supervision (ablation) | HumanGen texture generation · Texture generation | FID: 21.69 -> 20.97 | Helped |
| Add per-sample blending and second-stage image blending (ablation) | HumanGen texture generation · Texture generation | FID: 48.37 -> 20.97 | Helped |
| Add second-stage image blending (ablation) | HumanGen texture generation · Texture generation | FID: 25.94 -> 20.97 | Helped |
| Train with balanced view coverage, including back views (ablation) | HumanGen · Texture generation | FID: 21.81 -> 20.97 | Helped |
| Use the pretrained StyleGAN-human mapping network (ablation) | HumanGen texture fitting · Texture fitting | PSNR: 15.86 -> 21.13 | Helped |
| Add geometry embedding to the decoder (ablation) | HumanGen texture fitting · Texture fitting | PSNR: 19.89 -> 21.13 | Helped |
| Use a conditioned discriminator and the full adversarial-training setup (ablation) | HumanGen texture generation · Texture generation | FID: 31.07 -> 20.97 | Helped |
| Use the full HumanGen geometry and appearance design | Human generation · Human generation | Depth: EG3D 0.0395; StyleSDF 0.0493; 2D-G + PIFu 0.0379; GNARF 0.0408; Ours 0.0201 | Helped |
Closest papers by meaning
- HumanGaussian: Text-Driven 3D Human Generation with Gaussian Splatting (CVPR 2024 · similarity 0.88)
- En3D: An Enhanced Generative Model for Sculpting 3D Humans from 2D Synthetic Data (CVPR 2024 · similarity 0.87)
- GaussianIP: Identity-Preserving Realistic 3D Human Generation via Human-Centric Diffusion Prior (CVPR 2025 · similarity 0.87)
- HyperHuman: Hyper-Realistic Human Generation with Latent Structural Diffusion (ICLR 2024 · similarity 0.86)
- HumanNorm: Learning Normal Diffusion Model for High-quality and Realistic 3D Human Generation (CVPR 2024 · similarity 0.86)
- Joint2Human: High-Quality 3D Human Generation via Compact Spherical Embedding of 3D Joints (CVPR 2024 · 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.