A Comparative Study of Cross-Model Universal Adversarial Perturbation for Face Forgery | IEEE Conference Publication | IEEE Xplore