$\mu$Flow: Leveraging Average Images for Improving Generalisation of Deepfake Faces Detectors
arXiv:2606.30528v1 Announce Type: cross Abstract: Current generative models, including GANs and diffusion models, have reached an outstanding level of photorealism, posing significant risks to privacy and security. To ensure real-world applicability, deepfake detectors must generalise effectively to unseen generators. However, most existing approaches rely on supervised training with both real and fake images, which limits their generalisation especially across generators categories (e.g. GANs vs DMs). In this work, we introduce $\mu$Flow, a one-class deepfake detector trained only on real ima
The rapid advancement of generative AI models, especially GANs and diffusion models, has created an urgent need for robust deepfake detection methods that can generalize beyond their training data.
This development addresses a critical vulnerability in digital authenticity, impacting information integrity, cybersecurity, and societal trust in visual media.
The ability to detect deepfakes effectively across various generative models, even unseen ones, shifts the cat-and-mouse game towards more resilient defense mechanisms.
- · Cybersecurity firms
- · Social media platforms
- · Digital forensics
- · Journalism
- · Deepfake creators
- · Misinformation campaigns
Improved deepfake detection tools become more widely adopted across various digital platforms.
Reduced effectiveness of sophisticated deepfake attacks could lead to a decrease in their prevalence for certain use cases.
Enhanced trust in digital media, or conversely, a renewed focus on non-visual verification methods as deepfake generation also evolves.
This signal links to a primary source. Continuum Brief monitors and indexes it as part of the live intelligence stream — we do not republish source content.
Read at arXiv cs.LG