OmniMark: Efficient and Scalable Latent Diffusion Model Fingerprinting
We introduce OmniMark, a novel and efficient fingerprinting method for Latent Diffusion Models (LDM). OmniMark can encode user-specific fingerprints across diverse dimensions of the weights of the LDM, including kernels, filters, channels, and spatial domains. The LDM is fine-tuned to encode the invisible fingerprint into generated images, which can be decoded by a decoder. By altering fingerprints and re-encoding the weights, OmniMark supports efficient and scalable ad-hoc generation (
Authors
- Zhihua Xia (ORCID: https://orcid.org/0000-0001-6860-647X)
- Jianwei Fei (ORCID: https://orcid.org/0000-0002-1243-3909)
- Jianhui Zhou (ORCID: https://orcid.org/0000-0001-7293-9787)
- Fangjun Huang (ORCID: https://orcid.org/0000-0002-3098-3373)
- Yunshu Dai
Institutions
- Sun Yat-sen University (CN)
- University of Macau (MO)
- DigitalSpace (United States) (US)
- City University of Macau (MO)
Publication Details
- Journal
- Proceedings of the AAAI Conference on Artificial Intelligence
- Published
- 2025-04-11
- DOI
- https://doi.org/10.1609/aaai.v39i16.33818
- Citations
- 2
- Primary Topic
- Speech Recognition and Synthesis
- Type
- article
- Field-Weighted Citation Impact
- 1.77
Funders
- United Mitochondrial Disease Foundation
- National Natural Science Foundation of China
- Natural Science Foundation of Guangdong Province
- Universidade de Macau
- National Key Research and Development Program of China