Synthetic Face Datasets Generation via Latent Space Exploration from Brownian Identity Diffusion
This work addresses privacy and ethical concerns in face recognition by providing a method to generate synthetic datasets, though it appears incremental as it builds on existing generative approaches.
The authors tackled the problem of generating diverse synthetic face datasets for training face recognition models by introducing a method based on Brownian motion in latent space, which achieved competitive performance with state-of-the-art diffusion-based datasets and prevented data leakage.
Face recognition models are trained on large-scale datasets, which have privacy and ethical concerns. Lately, the use of synthetic data to complement or replace genuine data for the training of face recognition models has been proposed. While promising results have been obtained, it still remains unclear if generative models can yield diverse enough data for such tasks. In this work, we introduce a new method, inspired by the physical motion of soft particles subjected to stochastic Brownian forces, allowing us to sample identities distributions in a latent space under various constraints. We introduce three complementary algorithms, called Langevin, Dispersion, and DisCo, aimed at generating large synthetic face datasets. With this in hands, we generate several face datasets and benchmark them by training face recognition models, showing that data generated with our method exceeds the performance of previously GAN-based datasets and achieves competitive performance with state-of-the-art diffusion-based synthetic datasets. While diffusion models are shown to memorize training data, we prevent leakage in our new synthetic datasets, paving the way for more responsible synthetic datasets.