Learning the Manifold of Authenticity: Hybrid-Curvature Representation Learning for Generalizable Deepfake Detection

The practical utility of deepfake detectors is crippled by a crisis of generalization: models that perform well on known manipulation techniques consistently fail when faced with unseen forgeries. We argue this failure stems from a fundamental geometric mismatch. Existing methods implicitly assume t...

Popoln opis

Bibliografske podrobnosti

Principais autores:	Nicolas Larue, Vitomir Struc, Peter Peer, Ngoc-Son Vu
Format:	Artigo
Jezik:	Inglês
Izdano:	IEEE 2026-01-01
Serija:	IEEE Access
Teme:	Contrastive learning generalized deepfake detection hyperbolic one-class anomaly detection
Online dostop:	https://ieeexplore.ieee.org/document/11557307/
Oznake:	Označite Brez oznak, prvi označite!