Research map / Image generation / Text-to-image diffusion
StyleGAN-Human: A Data-Centric Odyssey of Human Generation
Jianglin Fu, Shikai Li, Yuming Jiang, Kwan-Yee Lin, Chen Qian, Chen Change Loy, Wayne Wu, Ziwei Liu
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 |
|---|---|---|---|
| increase training data from 10K to 20K (ablation) | StyleGAN2 · SHHQ 512x256 | FID: 7.80 -> 4.46 | Improved |
| increase training data from 10K to 20K (ablation) | StyleGAN2 · SHHQ 1024x512 | FID: 7.23 -> 4.33 | Improved |
| increase training data from 20K to 40K (ablation) | StyleGAN2 · SHHQ 1024x512 | FID: 4.33 -> 2.80 | Improved |
| increase training data from 40K to 230K (ablation) | StyleGAN2 · SHHQ 1024x512 | FID: 2.80 -> 1.57 | Improved |
| increase training data from 40K to 230K (ablation) | StyleGAN2 · SHHQ 512x256 | FID: 2.61 -> 1.97 | Improved |
| increase training data from 10K to 230K (ablation) | StyleGAN2 · SHHQ 1024x512 | IS: 3.93 -> 5.02 | Improved |
| add adaptive discriminator augmentation on small subsets (ablation) | StyleGAN2 · SHHQ S0-S2 | FID | Improved |
| use uniform instead of long-tailed face-yaw distribution (ablation) | StyleGAN2 · SHHQ face crops FID per yaw bin | FID | Improved |
| use uniform instead of long-tailed clothing texture distribution (ablation) | StyleGAN2 · SHHQ clothing patches FID per bin | FID | Mixed |
| align by pelvis instead of face center (ablation) | StyleGAN2 · SHHQ 50K alignment | FID: 3.5 -> 2.8 | Improved |
| align by mid-body instead of pelvis (ablation) | StyleGAN2 · SHHQ 50K alignment | FID: 2.8 -> 2.4 | Improved |
Closest papers by meaning
- UnitedHuman: Harnessing Multi-Source Data for High-Resolution Human Generation (ICCV 2023 · similarity 0.91)
- HyperHuman: Hyper-Realistic Human Generation with Latent Structural Diffusion (ICLR 2024 · similarity 0.88)
- OpenHumanVid: A Large-Scale High-Quality Dataset for Enhancing Human-Centric Video Generation (CVPR 2025 · similarity 0.87)
- CapHuman: Capture Your Moments in Parallel Universes (CVPR 2024 · similarity 0.87)
- MultiHuman-Testbench: Benchmarking Image Generation for Multiple Humans (NeurIPS 2025 · similarity 0.87)
- RealisHuman: A Two-Stage Approach for Refining Malformed Human Parts in Generated Images (AAAI 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.
