{"ID":2873509,"CreatedAt":"2026-06-01T04:54:23.091178241Z","UpdatedAt":"2026-06-01T04:54:23.091178241Z","DeletedAt":null,"paper_url":"https://arxiv.org/abs/2509.06818","arxiv_id":"2509.06818","title":"UMO: Scaling Multi-Identity Consistency for Image Customization via Matching Reward","abstract":"Recent advancements in image customization exhibit a wide range of application prospects due to stronger customization capabilities. However, since we humans are more sensitive to faces, a significant challenge remains in preserving consistent identity while avoiding identity confusion with multi-reference images, limiting the identity scalability of customization models. To address this, we present UMO, a Unified Multi-identity Optimization framework, designed to maintain high-fidelity identity preservation and alleviate identity confusion with scalability. With \"multi-to-multi matching\" paradigm, UMO reformulates multi-identity generation as a global assignment optimization problem and unleashes multi-identity consistency for existing image customization methods generally through reinforcement learning on diffusion models. To facilitate the training of UMO, we develop a scalable customization dataset with multi-reference images, consisting of both synthesised and real parts. Additionally, we propose a new metric to measure identity confusion. Extensive experiments demonstrate that UMO not only improves identity consistency significantly, but also reduces identity confusion on several image customization methods, setting a new state-of-the-art among open-source methods along the dimension of identity preserving. Code and model: https://github.com/bytedance/UMO","short_abstract":"Recent advancements in image customization exhibit a wide range of application prospects due to stronger customization capabilities. However, since we humans are more sensitive to faces, a significant challenge remains in preserving consistent identity while avoiding identity confusion with multi-reference images, limi...","url_abs":"https://arxiv.org/abs/2509.06818","url_pdf":"https://arxiv.org/pdf/2509.06818v1","authors":"[\"Yufeng Cheng\",\"Wenxu Wu\",\"Shaojin Wu\",\"Mengqi Huang\",\"Fei Ding\",\"Qian He\"]","published":"2025-09-08T15:54:55Z","proceeding":"cs.CV","tasks":"[\"cs.CV\",\"cs.LG\"]","methods":"[\"Reinforcement Learning\",\"Diffusion Model\"]","has_code":false,"code_links":[{"ID":610060,"CreatedAt":"2026-06-01T04:54:23.091178241Z","UpdatedAt":"2026-06-01T04:54:23.091178241Z","DeletedAt":null,"paper_id":2873509,"paper_url":"https://arxiv.org/abs/2509.06818","paper_title":"UMO: Scaling Multi-Identity Consistency for Image Customization via Matching Reward","repo_url":"https://github.com/bytedance/UMO","is_official":false,"mentioned_in_paper":false,"mentioned_in_github":true,"github_stars":0}]}