Ee-Chien Chang
Titles as the sources gave them. In 1 entry, with 0 checked quotations.
In the Base
Every entry they appear in.
Provably Secure Diffusion Watermarking Against Removal Adversaries with Batched Detector Access
According to the abstract, De Zhang Lee and Ee-Chien Chang show that existing latent watermarks for AI-generated images can be broken by an attacker who queries a public detector in batches. They propose a scheme combining long-term and one-time secrets and claim a security proof, with tests on Stable Diffusion 2.1 and XL.