Reproduce “Towards General Visual-Linguistic Face Forgery Detection” on OpenResearch
Turn your favorite coding agents into research agents that review literature, develop hypotheses, and run experiments.
Universal app for Apple silicon and Intel. Requires macOS 11 or later.
About this paper
Ke Sun, Shen Chen, Taiping Yao, Haozhe Yang, Xiaoshuai Sun, Shouhong Ding, Rongrong Ji
VLFFD trains a visual-and-language model for face-forgery detection using real/fake labels alongside generated mixed-face examples paired with region and forgery-type descriptions. Its coarse and fine objectives are trained together, supporting binary classification and descriptive outputs. Under FaceForensics++ high-quality training and unseen DFDC-P testing, it reaches 84.74% frame-level AUC versus DCL’s 76.71%; prompts are heuristic, not verified explanations.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.