ID-Booth: Identity-consistent image generation with diffusion models

Darian Tomašević, Fadi Boutros, Naser Damer, Vitomir Struc, Peter Peer

Keywords: Image synthesis, Diffusion models, Face recognition data

Abstract: The recent retraction of large-scale biometric datasets, prompted by strict privacy regulations, presents a critical challenge for future biometric research. This is evident with the face recognition task, for which large-scale datasets were often gathered through web-scraping without the consent of subjects. A potential solution entails the creation of synthetic data, suitable for training recognition models, with deep generative models. Existing generative approaches rely on conditioning and fine-tuning of powerful pretrained diffusion models to achieve the synthesis of realistic images of a desired identity. Yet, these methods often do not consider the identity of subjects during training, leading to poor consistency between generated and intended identities. In contrast, methods that employ identity-based training objectives tend to overfit on various aspects of the identity, and in turn, lower the diversity of images that can be generated. To address these issues, we present the ID-Booth fine-tuning framework, which utilizes a novel triplet identity training objective and enables identity-consistent image generation while retaining the synthesis capabilities of pretrained models. Experiments across two latent diffusion models with varying prompt complexity reveal that our method facilitates better intra-identity consistency and inter-identity separability while achieving higher image diversity. In turn, the produced data enables the training of better-performing recognition models than even real-world datasets of a similar scale gathered with suitable consent. The source code for the ID-Booth framework is available at omitted_for_review.

Primary Area: generative models

Code Of Ethics: I acknowledge that I and all co-authors of this work have read and commit to adhering to the ICLR Code of Ethics.

Submission Guidelines: I certify that this submission complies with the submission instructions as described on https://iclr.cc/Conferences/2025/AuthorGuide.

Reciprocal Reviewing: I understand the reciprocal reviewing requirement as described on https://iclr.cc/Conferences/2025/CallForPapers. If none of the authors are registered as a reviewer, it may result in a desk rejection at the discretion of the program chairs. To request an exception, please complete this form at https://forms.gle/Huojr6VjkFxiQsUp6.

Anonymous Url: I certify that there is no URL (e.g., github page) that could be used to find authors’ identity.

No Acknowledgement Section: I certify that there is no acknowledgement section in this submission for double blind review.

Submission Number: 13546