We had to make images fit for Deepfake production by both face enhancement and colorization. It was impossible with existing algorithms, so we developed our own method in Phygital+, and so could make more images with different emotional expressions and camera angles. After tests and iterations, only one image turned out to be a good fit for face swapping