Pulsar: Secure Steganography through Diffusion Models

ePrint Report: Pulsar: Secure Steganography through Diffusion Models

Tushar M. Jois, Gabrielle Beck, Gabriel Kaptchuk

Widespread efforts to subvert acccess to strong cryptography has renewed interest in steganography, the practice of embedding sensitive messages in mundane cover messages. Recent efforts at provably secure steganography have only focused on text-based generative models and cannot support other types of models, such as diffusion models, which are used for high-quality image synthesis. In this work, we initiate the study of securely embedding steganographic …

cryptography diffusion models eprint report generative generative models interest messages practice pulsar report sensitive steganography support text