Visual Analysis & Security (VAS) Lab conducts research on Trustworthy Visual Intelligence, with particular interests in multimedia forensics, vision security, and machine learning. Our mission is to analyze, authenticate, and protect visual content against manipulation, misinformation, and malicious misuse, while advancing the theoretical foundations that improve the robustness, security, and trustworthiness of modern AI-driven visual systems. Our research is organized around three major directions:
- Multimedia Forensics. Developing advanced forensic algorithms for images, videos, and audio to detect, localize, and analyze manipulated or AI-generated content, covering both passive and active defense in both general-purpose and face-centric scenarios.
- Vision Algorithms. Studying computer vision and machine learning algorithms for visual understanding and modeling, aiming to support the development of robust, reliable, and secure visual intelligence systems.
- Vision Security. Investigating the vulnerabilities of vision models, including adversarial and backdoor attacks, data security, and model privacy, as well as secure deployment in real-world applications, to improve the robustness, reliability, and security of visual intelligence systems.
We are always looking for motivated students who are passionate about research and emerging technologies, especially those with strong interests in mathematics and programming, to join our lab. If you are interested, please drop me an email with your CV at liyuezun [AT] ouc.edu.cn.
🔉 Optimism is a happiness magnet — Mary Lou Retton