Zhongjie Ba

Attack-Resistant Watermarking for AIGC Image Forensics via Diffusion-based Semantic Deflection

The Robustness-Security Paradox: Channel-Aware Feature Learning for Adversarial Watermark Exploitation

WMCopier: Forging Invisible Watermarks on Arbitrary Images

Robust Representation Consistency Model via Contrastive Denoising

JudgeRail: Harnessing Open-Source LLMs for Fast Harmful Text Detection with Judicial Prompting and Logit Rectification

A Comprehensive Deepfake Detector Assessment Platform