Skip to content

Latest commit

 

History

History
18 lines (15 loc) · 760 Bytes

README.md

File metadata and controls

18 lines (15 loc) · 760 Bytes

Video Forgery Detection

Detection of Object-Based Forgery in Advanced Video based on this paper

Dependencies

  1. ffmpeg
  2. MATLAB
  3. Video dataset for training and testing (Available here)
  4. JPEG toolbox (Included)
  5. CC-PEV 548 (Included)
  6. Ensemble Classifier (Included)

Steps

  1. Extract frames from a video using extract_frames.sh
  2. Use main_residuals.m to extract motion residuals from frames
  3. Use extract_features.m to extract CC-PEV features from motion residuals
  4. Use train.m to train ensemble classifier with given features
  5. Use test.m to test the classifier on new datasets