A Black-Box Watermarking Modulation for Object Detection Models

Authors

  • Mohammed Lansari Thales SIX GTS, 1 Av. Augustin Fresnel, 91120 Palaiseau, France IMT Atlantique, 655 Av. du Technopôle, 29280 Plouzané, France
  • Lucas Mattioli IRT SystemX, 2 Boulevard Thomas Gobert, 91120 Palaiseau, France
  • Boussad Addad Thales SIX GTS, 1 Av. Augustin Fresnel, 91120 Palaiseau, France IRT SystemX, 2 Boulevard Thomas Gobert, 91120 Palaiseau, France
  • Paul-Marie Raffi IRT SystemX, 2 Boulevard Thomas Gobert, 91120 Palaiseau, France
  • Katarzyna Kapusta Thales SIX GTS, 1 Av. Augustin Fresnel, 91120 Palaiseau, France IRT SystemX, 2 Boulevard Thomas Gobert, 91120 Palaiseau, France
  • Martin Gonzalez IRT SystemX, 2 Boulevard Thomas Gobert, 91120 Palaiseau, France
  • Mohamed Ibn Khedher IRT SystemX, 2 Boulevard Thomas Gobert, 91120 Palaiseau, France

DOI:

https://doi.org/10.1609/aaaiss.v4i1.31772

Abstract

Training a Deep Neural Network (DNN) from scratch comes with a substantial cost in terms of money, energy, data, and hardware. When such models are misused or redistributed without authorisation, the owner faces significant financial and intellectual property (IP) losses. Therefore, there is a pressing need to protect the IP of Machine Learning models to avoid these issues. ML watermarking emerges as a promising solution for model traceability. Watermarking has been well-studied for image classification models, but there is a significant research gap in its application to other tasks like object detection, for which no effective methods have been proposed yet. In this paper, we introduce a novel black-box watermarking method for object detection models. Our contributions include a watermarking technique that maps visual information to text semantics and a comparative study of fine-tuning techniques’ impact on watermark detectability. We present the model’s detection performance and evaluate fine-tuning strategies’ effectiveness in preserving watermark integrity.

Downloads

Published

2024-11-08

How to Cite

Lansari, M., Mattioli, L., Addad, B., Raffi, P.-M., Kapusta, K., Gonzalez, M., & Ibn Khedher, M. (2024). A Black-Box Watermarking Modulation for Object Detection Models. Proceedings of the AAAI Symposium Series, 4(1), 60-67. https://doi.org/10.1609/aaaiss.v4i1.31772

Issue

Section

AI Trustworthiness and Risk Assessment for Challenging Contexts (ATRACC)