EUVIMEDEUVIMED entdecken
Crossref

FocusDPO: Dynamic Preference Optimization for Multi-Subject Personalized Image Generation via Adaptive Focus

Qiaoqiao Jin, Siming Fu, Dong She, Weinan Jia, Hualiang Wang, Mu Liu, Jidong Jiang

Proceedings of the AAAI Conference on Artificial Intelligence · 14.03.2026 · 40 (7) · 5512-5520

DOI: 10.1609/aaai.v40i7.37469 ↗

Kurz gefasst

Fachartikel in der Zeitschrift Proceedings of the AAAI Conference on Artificial Intelligence, online erschienen am 14.03.2026. Verfasst von Qiaoqiao Jin und 6 weiteren Autor:innen.

AbstractBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

Multi-subject personalized image generation aims to synthesize customized images containing multiple specified subjects without requiring test-time optimization. However, achieving fine-grained independent control over multiple subjects remains challenging due to difficulties in preserving subject fidelity and preventing cross-subject attribute leakage. We present FocusDPO, a framework that adaptively identifies focus regions based on dynamic semantic correspondence and supervision image complexity. During training, our method progressively adjusts these focal areas across noise timesteps, implementing a weighted strategy that rewards information-rich patches while penalizing regions with low prediction confidence. The framework dynamically adjusts focus allocation during the DPO process according to the semantic complexity of reference images and establishes robust correspondence mappings between generated and reference subjects. Extensive experiments demonstrate that our method substantially enhances the performance of existing pre-trained personalized generation models, achieving state-of-the-art results on both single-subject and multi-subject personalized image synthesis benchmarks. Our method effectively mitigates attribute leakage while preserving superior subject fidelity across diverse generation scenarios, advancing the frontier of controllable multi-subject image synthesis.

Zitieren

APA

Jin, Q., Fu, S., She, D., Jia, W., Wang, H., Liu, M., & Jiang, J. (2026). FocusDPO: Dynamic Preference Optimization for Multi-Subject Personalized Image Generation via Adaptive Focus. Proceedings of the AAAI Conference on Artificial Intelligence, 40(7), 5512-5520. https://doi.org/10.1609/aaai.v40i7.37469

Vancouver

Jin Q, Fu S, She D, Jia W, Wang H, Liu M, et al. FocusDPO: Dynamic Preference Optimization for Multi-Subject Personalized Image Generation via Adaptive Focus. Proceedings of the AAAI Conference on Artificial Intelligence. 2026;40(7):5512-5520. doi:10.1609/aaai.v40i7.37469

BibTeX
@article{Jin2026,
  author = {Jin, Qiaoqiao and Fu, Siming and She, Dong and Jia, Weinan and Wang, Hualiang and Liu, Mu and Jiang, Jidong},
  title = {FocusDPO: Dynamic Preference Optimization for Multi-Subject Personalized Image Generation via Adaptive Focus},
  journal = {Proceedings of the AAAI Conference on Artificial Intelligence},
  year = {2026},
  volume = {40},
  number = {7},
  pages = {5512--5520},
  doi = {10.1609/aaai.v40i7.37469}
}
Kostenlos starten →