/video-watermarking-forgery-detection

Robustly Hiding Mutiple images into a single image

Primary LanguagePython

Code for "Robust Watermarking for Video Forgery Detection with Improved Imperceptibility and Robustness"

https://arxiv.org/pdf/2207.03409 MMSP 2022 Poster

Dataset

Our contributions are mainly as follows:

  1. We use deep networks for video watermarking against tampering;
  2. We propose a tailored attacking layer for enhanced robustness against typical video post-processing attacks;
  3. Our method can achieve higher accuracy in locating tampered regions and is robust against multiple kinds of post-processing attacks.