Multimedia Forensics: Challenges and Opportunities in the Era of Large Models - Livestream

With the implementation and wide application of related tools such as ChatGPT and Sora, people realized that seeing is indeed no longer believing. More recently, the emergence of diffusion models with powerful generation capabilities has driven this topic forward. Subsequently, as the powerful tools for deepfake generation, corresponding detection technologies continuously evolved hence raising more security concerns of the public. This talk will discuss the development of multimedia forensics and focus on researching four representative deepfake fields: face swapping/face reenactment, talking face generation, and facial attribute editing, as well as forgery and deepfake detection and will finally analyze thechallenges and future research directions of the discussed fields.

Register at weblink