PulseExploreJournal ClubDebatesTrendingResearchersJournals
Instagram
HomeExploreJournal ClubTrending
Synapse
⌘+K
Synapse
June 1, 2022163 citations

Leveraging Real Talking Faces via Self-Supervision for Robust Forgery Detection

View Full Paper
AHAlexandros HaliassosRMRodrigo MiraSPStavros Petridis

Key Points

Key points are not available for this paper at this time.

Abstract

One of the most pressing challenges for the detection of face-manipulated videos is generalising to forgery methods not seen during training while remaining effective under common corruptions such as compression. In this paper, we examine whether we can tackle this issue by harnessing videos of real talking faces, which contain rich information on natural facial appearance and behaviour and are readily available in large quantities online. Our method, termed RealForensics, consists of two stages. First, we exploit the natural correspondence between the visual and auditory modalities in real videos to learn, in a self-supervised cross-modal manner, temporally dense video representations that capture factors such as facial movements, expression, and identity. Second, we use these learned representations as targets to be predicted by our forgery detector along with the usual binary forgery classification task; this encourages it to base its real/fake decision on said factors. We show that our method achieves state-of-the-art performance on cross-manipulation generalisation and robustness experiments, and examine the factors that contribute to its per-formance. Our results suggest that leveraging natural and unlabelled videos is a promising direction for the development of more robust face forgery detectors.

Ask AI
Helpful
Bookmark
Share
View Full Paper

Cite This Study

Haliassos et al. (2022) studied this question.

synapsesocial.com/papers/69da44ce0f778bd2e468499bhttps://doi.org/10.1109/cvpr52688.2022.01453
Ask AI
Helpful
Bookmark
Share
View Full Paper