CycleGAN

CycleGAN is a type of generative adversarial network designed for unpaired image-to-image translation, enabling conversion between visual domains without requiring matched training examples.

Referenced by (1)

Full triples — surface form annotated when it differs from this entity's canonical label.