The field of media forensics and deepfake detection is rapidly evolving, with a focus on developing innovative methods to identify and mitigate manipulated media. Recent research has explored the use of multimodal approaches, combining visual and audio cues to detect deepfakes. Additionally, there is a growing interest in developing explainable and transparent models that can provide insights into the decision-making process. The use of large language models and contrastive learning techniques has also shown promise in improving the accuracy and robustness of deepfake detection systems. Notably, papers such as MER-CLIP and FauForensics have proposed novel approaches to micro-expression recognition and audio-visual deepfake detection, respectively. Furthermore, the development of new datasets, such as TT-DF, and benchmarks, such as DFA-CON, will facilitate further research and advancements in this field.
Advances in Media Forensics and Deepfake Detection
Sources
MER-CLIP: AU-Guided Vision-Language Alignment for Micro-Expression Recognition
Unmasking Deep Fakes: Leveraging Deep Learning for Video Authenticity Detection
Minimizing Risk Through Minimizing Model-Data Interaction: A Protocol For Relying on Proxy Tasks When Designing Child Sexual Abuse Imagery Detection Models
Beyond Identity: A Generalizable Approach for Deepfake Audio Detection
Multimodal Fake News Detection: MFND Dataset and Shallow-Deep Multitask Learning
DP-TRAE: A Dual-Phase Merging Transferable Reversible Adversarial Example for Image Privacy Protection
MELLM: Exploring LLM-Powered Micro-Expression Understanding Enhanced by Subtle Motion Perception
MarkMatch: Same-Hand Stuffing Detection
LAMM-ViT: AI Face Detection via Layer-Aware Modulation of Region-Guided Attention
Visual Watermarking in the Era of Diffusion Models: Advances and Challenges
Where the Devil Hides: Deepfake Detectors Can No Longer Be Trusted
Disruptive Transformation of Artworks in Master-Disciple Relationships: The Case of Ukiyo-e Artworks
FauForensics: Boosting Audio-Visual Deepfake Detection with Facial Action Units
TT-DF: A Large-Scale Diffusion-Based Dataset and Benchmark for Human Body Forgery Detection
DFA-CON: A Contrastive Learning Approach for Detecting Copyright Infringement in DeepFake Art
WaveGuard: Robust Deepfake Detection and Source Tracing via Dual-Tree Complex Wavelet and Graph Neural Networks
MixBridge: Heterogeneous Image-to-Image Backdoor Attack through Mixture of Schr\"odinger Bridges
FaceShield: Explainable Face Anti-Spoofing with Multimodal Large Language Models
Denoising and Alignment: Rethinking Domain Generalization for Multimodal Face Anti-Spoofing
MorphGuard: Morph Specific Margin Loss for Enhancing Robustness to Face Morphing Attacks