Research map / 3D vision / Face and shape
Fake It Till You Make It: Face Analysis in the Wild Using Synthetic Data Alone
Erroll Wood, Tadas Baltrušaitis, Charlie Hewitt, Sebastian Dziadzio, Thomas J. Cashman, Jamie Shotton
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 full augmentation (appearance and geometry) to synthetic training (ablation) | Ours (synthetic) · 300W Common | NME: 4.25 -> 3.09 | Improved |
| add geometric augmentation on top of appearance augmentation (ablation) | Ours (synthetic) · 300W Challenging | NME: 6.80 -> 4.86 | Improved |
| add label adaptation to synthetic-trained model (ablation) | Ours (synthetic) · 300W Common | NME: 5.61 -> 3.09 | Improved |
| add label adaptation to synthetic-trained model (ablation) | Ours (synthetic) · 300W Private | FR10%: 4.67 -> 0.50 | Improved |
| add label adaptation to real-data-trained model (ablation) | Ours (real) · 300W Common | NME: 3.44 -> 3.37 | No clear effect |
| add hair and clothing to synthetic faces (ablation) | Ours (synthetic) · 300W Common | NME: 3.36 -> 3.09 | Improved |
| add hair and clothing to synthetic faces (ablation) | Ours (synthetic) · 300W Private | FR10%: 2.17 -> 0.50 | Improved |
| add clothing to synthetic faces (ablation) | Ours (synthetic) · 300W Challenging | NME: 5.09 -> 4.86 | Mixed |
| add hair on top of clothing-free faces (ablation) | Ours (synthetic) · 300W Challenging | NME: 5.37 -> 5.09 | Mixed |
| train on synthetic instead of real data | landmark localization ResNet34 · 300W Challenging | NME: 5.77 -> 4.86 | Improved |
| train on synthetic instead of real data | face parsing UNet · Helen* face parsing (overall) | F1: 91.6 -> 92.0 | Improved |
| train on synthetic instead of real data | face parsing UNet · LaPa face parsing (mean) | F1: 90.9 -> 90.1 | Hurt |
Closest papers by meaning
- HyperFace: Generating Synthetic Face Recognition Datasets by Exploring Face Embedding Hypersphere (ICLR 2025 · similarity 0.89)
- UIFace: Unleashing Inherent Model Capabilities to Enhance Intra-Class Diversity in Synthetic Face Recognition (ICLR 2025 · similarity 0.87)
- Synthetic Face Datasets Generation via Latent Space Exploration from Brownian Identity Diffusion (ICML 2025 · similarity 0.87)
- $\text{ID}^3$: Identity-Preserving-yet-Diversified Diffusion Models for Synthetic Face Recognition (NeurIPS 2024 · similarity 0.87)
- Real-Fake: Effective Training Data Synthesis Through Distribution Matching (ICLR 2024 · similarity 0.87)
- Data Synthesis with Diverse Styles for Face Recognition via 3DMM-Guided Diffusion (CVPR 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.
