IPRemover: A Generative Model Inversion Attack against Deep Neural Network Fingerprinting and Watermarking.
Wei ZongYang-Wai ChowWilly SusiloJoonsang BaekJongkil KimSeyit CamtepePublished in: AAAI (2024)
Keyphrases
- generative model
- neural network
- copyright protection
- digital watermark
- probabilistic model
- bayesian framework
- posterior probability
- digital images
- watermarking scheme
- discriminative learning
- em algorithm
- latent dirichlet allocation
- semi supervised
- discriminative models
- prior knowledge
- topic models
- watermarking algorithm
- learned models
- digital content
- multiscale
- watermarking technique
- multimedia content
- markov chain monte carlo
- low level
- image reconstruction
- video representation
- image processing