CtlGAN introduces a novel contrastive transfer learning strategy for generating high-quality artistic portraits from real face photos with no more than 10 training examples.
Proposing CtlGAN for few-shot artistic portraits generation with contrastive transfer learning to prevent overfitting and ensure high-quality results.
The author proposes CtlGAN, a new model for few-shot artistic portraits generation, utilizing contrastive transfer learning and a novel encoder approach to achieve high-quality results.