Different kind of training for CR7 to use him as a virtual avatar with Pix2Pix (using the code of datitran).
I generated the dataset from this video. The Network is trained with OpenCV landmark images from a source video. The same landmark detection is applied to the webcam output.