A Deep Learning classification to detect faked videos from video inputs and render them as an exported output with annotaions of people's faces in the video as real or fake, using FaceForensics++ paper @ ~60% accuracy on HD videos  

This is an implmentation of XceptionNet from the paper trained on our FaceForensics++ dataset. Besides the full image models, all models were trained on slightly enlarged face crops with a scale factor of 1.3. The models were trained using the Face2Face face tracker, though the detect_from_models.py file uses the freely available dlib face detector.

Note that we provide the trained models from our paper which have not been fine-tuned for general compressed videos. You can find our used models under this link.

$ pip install -r requirements.txt

Setup:

• Install required modules via requirement.txt file
• Run detection from a single video file or folder with

python detect_from_video.py
-i <path to input video or folder of videos with extenstion '.mp4' or '.avi'>
-m <path to model file, default is imagenet model
-o <path to output folder, will contain output video(s)

from the classification folder. Enable cuda with --cuda or see parameters with python detect_from_video.py -h.