Face Transfer with Generative Adversarial Network
This work addresses face animation for video production or entertainment, but it is incremental as it builds on existing GAN techniques like CycleGAN and PatchGAN.
The paper tackles the problem of animating a target character's face in video using a source actor's performance by proposing an end-to-end face transfer method based on Generative Adversarial Networks, specifically leveraging CycleGAN to generate images with matching head poses and expressions, and exploring PatchGAN with different receptive fields to improve video quality.
Face transfer animates the facial performances of the character in the target video by a source actor. Traditional methods are typically based on face modeling. We propose an end-to-end face transfer method based on Generative Adversarial Network. Specifically, we leverage CycleGAN to generate the face image of the target character with the corresponding head pose and facial expression of the source. In order to improve the quality of generated videos, we adopt PatchGAN and explore the effect of different receptive field sizes on generated images.