The field of deepfake detection and media authentication is rapidly evolving, with a focus on developing innovative methods to combat the growing threat of synthetic media. Recent research has explored various approaches, including facial feature extraction, handcrafted frequency-domain features, and audio watermarking. These advancements aim to improve the accuracy and robustness of deepfake detection systems, particularly in real-world scenarios. Noteworthy papers in this area include 'Deepfake Detection Via Facial Feature Extraction and Modeling', which introduces a novel approach using facial landmarks for deepfake detection, and 'WaveVerify: A Novel Audio Watermarking Framework for Media Authentication and Combatting Deepfakes', which proposes a framework for audio content authentication. Overall, the field is moving towards developing more effective and reliable methods for detecting and mitigating the risks associated with deepfakes and synthetic media.
Advances in Deepfake Detection and Media Authentication
Sources
How good are humans at detecting AI-generated images? Learnings from an experiment
Deepfake Detection Via Facial Feature Extraction and Modeling
Technological Requirements for Videoconferencing Judicial Hearings: Enhancing the Credibility and Reliability of Remote Testimonies
AI-Driven Media & Synthetic Knowledge: Rethinking Society in Generative Futures
HumanSAM: Classifying Human-centric Forgery Videos in Human Spatial, Appearance, and Motion Anomaly
Second Competition on Presentation Attack Detection on ID Card
Enhanced Deep Learning DeepFake Detection Integrating Handcrafted Features
FantasyID: A dataset for detecting digital manipulations of ID-documents
WaveVerify: A Novel Audio Watermarking Framework for Media Authentication and Combatting Deepfakes
Unmasking Synthetic Realities in Generative AI: A Comprehensive Review of Adversarially Robust Deepfake Detection Systems
Suppressing Gradient Conflict for Generalizable Deepfake Detection
Evaluating Deepfake Detectors in the Wild
Robust Deepfake Detection for Electronic Know Your Customer Systems Using Registered Images
Noise-Coded Illumination for Forensic and Photometric Video Analysis
X-NeMo: Expressive Neural Motion Reenactment via Disentangled Latent Attention
Digital literacy interventions can boost humans in discerning deepfakes