Wide Flat Minimum Watermarking for Robust Ownership Verification of GANs
- 14 August 2024
- journal article
- research article
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Information Forensics and Security
- Vol. 19 (15566013) , 8322-8337
Abstract
We propose a novel multi-bit box-free watermarking method for the protection of Intellectual Property Rights (IPR) of GANs with improved robustness against white-box model-level attacks like fine-tuning, pruning, quantization, and surrogate model attacks. The watermark is embedded by adding an extra watermarking loss term during GAN training, ensuring that the images generated by the GAN contain an invisible watermark that can be retrieved by a pre-trained watermark decoder. In order to improve the robustness against white-box model-level attacks, we make sure that the model converges to a wide flat minimum of the watermarking loss term, in such a way that any modification of the model parameters does not erase the watermark. To do so, we add random noise vectors to the parameters of the generator and require that the watermarking loss term is as invariant as possible with respect to the presence of noise. This procedure forces the generator to converge to a wide flat minimum of the watermarking loss. The proposed method is architecture- and dataset-agnostic, thus being applicable to many different generation tasks and models, as well as to CNN-based image processing architectures. We present the results of extensive experiments showing that the presence of the watermark has a negligible impact on the quality of the generated images, and proving the superior robustness of the watermark against model modification and surrogate model attacks.Keywords
All Related Versions
Funding Information
- National Key Research and Development Program of China (2022YFB3103100)
- National Natural Science Foundation of China (62122032, U23B2023)
This publication has 54 references indexed in Scilit:
- A novel model watermarking for protecting generative adversarial networkComputers & Security, 2023
- Intellectual property protection for deep semantic segmentation modelsFrontiers of Computer Science, 2022
- Spread-Transform Dither Modulation Watermarking of Deep Neural NetworkJournal of Information Security and Applications, 2021
- A survey of Deep Neural Network watermarking techniquesNeurocomputing, 2021
- Yes We can: Watermarking Machine Learning Models beyond Classification2021
- Protecting IP of Deep Neural Networks with Watermarking: A New Label Helps2020
- Adversarial frontier stitching for remote neural network watermarkingNeural Computing and Applications, 2019
- HiDDeN: Hiding Data With Deep Networks2018
- Digital watermarking for deep neural networksInternational Journal of Multimedia Information Retrieval, 2018
- The Pascal Visual Object Classes Challenge: A RetrospectiveInternational Journal of Computer Vision, 2014