# SteerFace: Debiasing Synthetic Face Generation via Adaptive Residue Perturbation

arXiv: https://arxiv.org/abs/2605.30894 · HTML: https://mlsanity.com/paper/2605.30894/

## Key facts

- No citing papers in the map yet (data through 14 Oct 2026).
- Builds on 37 of its 58 references that are in the map.
- First posted on arXiv 29 May 2026.

## Builds on (most cited first, 30 of 37)

- [Deep Residual Learning for Image Recognition](https://mlsanity.com/paper/1512.03385/) (2015)
- [Adam: A Method for Stochastic Optimization](https://mlsanity.com/paper/1412.6980/) (2014)
- [Denoising Diffusion Implicit Models](https://mlsanity.com/paper/2010.02502/) (2020)
- [A Style-Based Generator Architecture for Generative Adversarial Networks](https://mlsanity.com/paper/1812.04948/) (2018)
- [DreamBooth: Fine Tuning Text-to-Image Diffusion Models for Subject-Driven Generation](https://mlsanity.com/paper/2208.12242/) (2022)
- [An Image is Worth One Word: Personalizing Text-to-Image Generation using Textual Inversion](https://mlsanity.com/paper/2208.01618/) (2022)
- [CosFace: Large Margin Cosine Loss for Deep Face Recognition](https://mlsanity.com/paper/1801.09414/) (2018)
- [VGGFace2: A dataset for recognising faces across pose and age](https://mlsanity.com/paper/1710.08092/) (2017)
- [MS-Celeb-1M: A Dataset and Benchmark for Large-Scale Face Recognition](https://mlsanity.com/paper/1607.08221/) (2016)
- [Learning an Animatable Detailed 3D Face Model from In-The-Wild Images](https://mlsanity.com/paper/2012.04012/) (2020)
- [Learning Face Representation from Scratch](https://mlsanity.com/paper/1411.7923/) (2014)
- [AdaFace: Quality Adaptive Margin for Face Recognition](https://mlsanity.com/paper/2204.00964/) (2022)
- [PhotoMaker: Customizing Realistic Human Photos via Stacked ID Embedding](https://mlsanity.com/paper/2312.04461/) (2023)
- [FastComposer: Tuning-Free Multi-Subject Image Generation with Localized Attention](https://mlsanity.com/paper/2305.10431/) (2023)
- [Cross-Age LFW: A Database for Studying Cross-Age Face Recognition in Unconstrained Environments](https://mlsanity.com/paper/1708.08197/) (2017)
- [WebFace260M: A Benchmark Unveiling the Power of Million-Scale Deep Face Recognition](https://mlsanity.com/paper/2103.04098/) (2021)
- [LivePortrait: Efficient Portrait Animation with Stitching and Retargeting Control](https://mlsanity.com/paper/2407.03168/) (2024)
- [The MegaFace Benchmark: 1 Million Faces for Recognition at Scale](https://mlsanity.com/paper/1512.00596/) (2015)
- [DCFace: Synthetic Face Generation with Dual Condition Diffusion Model](https://mlsanity.com/paper/2304.07060/) (2023)
- [DigiFace-1M: 1 Million Digital Face Images for Face Recognition](https://mlsanity.com/paper/2210.02579/) (2022)
- [Face0: Instantaneously Conditioning a Text-to-Image Model on a Face](https://mlsanity.com/paper/2306.06638/) (2023)
- [DiffusionRig: Learning Personalized Priors for Facial Appearance Editing](https://mlsanity.com/paper/2304.06711/) (2023)
- [SDD-FIQA: Unsupervised Face Image Quality Assessment with Similarity Distribution Distance](https://mlsanity.com/paper/2103.05977/) (2021)
- [Inserting Anybody in Diffusion Models via Celeb Basis](https://mlsanity.com/paper/2306.00926/) (2023)
- [UV-GAN: Adversarial Facial UV Map Completion for Pose-invariant Face Recognition](https://mlsanity.com/paper/1712.04695/) (2017)
- [3D Guided Fine-Grained Face Manipulation](https://mlsanity.com/paper/1902.08900/) (2019)
- [Enhancing Detail Preservation for Customized Text-to-Image Generation: A Regularization-Free Approach](https://mlsanity.com/paper/2305.13579/) (2023)
- [StableIdentity: Inserting Anybody into Anywhere at First Sight](https://mlsanity.com/paper/2401.15975/) (2024)
- [MOST-GAN: 3D Morphable StyleGAN for Disentangled Face Image Manipulation](https://mlsanity.com/paper/2111.01048/) (2021)
- [Privacy-Preserving Face Recognition Using Trainable Feature Subtraction](https://mlsanity.com/paper/2403.12457/) (2024)

## Abstract (arXiv)

The shortage of legally compliant data for face recognition training has sparked growing interest in using synthetic data as an alternative. While recent diffusion-based methods enable the generation of photorealistic face images with strong identity adherence and data diversity, their downstream recognition performance still exhibits a significant synthetic-real gap. This paper identifies visual tendency as a previously underexplored limitation, whereby synthetic data exhibit an unrealistic prevalence of visual attributes and thus deviate from the real-data distribution. Visual tendency can be attributed to the generator's conditioning on identity embeddings, through which co-occurring residual visual cues are unintentionally absorbed into learned identity semantics. To discourage the generator from exploiting such visual cues, this paper proposes SteerFace, a simple and efficient training framework that perturbs identity embeddings by steering them toward random orthogonal directions on the embedding hypersphere. The perturbation serves as an identity-preserving regularizer that penalizes the generator's reliance on non-identity components, as supported by theoretical analysis. This paper further introduces an adaptive strategy that learns perturbation strengths with both sample-wise preference and favorable overall statistics. Extensive experiments show that SteerFace effectively mitigates visual tendency, outperforms prior methods in downstream face recognition, and generalizes well across different training datasets and generation pipelines.