1 paper · 1 filter
Ariel Elazary, Yotam Nitzan, Daniel Cohen-Or
In recent years, the role of image generative models in facial reenactment has been steadily increasing. Such models are usually subject-agnostic and trained on domain-wide dataset…