TransGaGa

Geometry-Aware Unsupervised Image-to-Image Translation