Few-Shot Image Generation via Cross-Domain Correspondence

Abstract

Training generative models, such as GANs, on a target domain containing limited examples (e.g., 10) can easily result in overfitting. In this work, we seek to utilize a large source domain for pretraining and transfer the diversity information from source to target. We propose to preserve the relative similarities and differences between instances in the source via a novel cross-domain distance consistency loss. To further reduce overfitting, we present an anchor-based strategy to encourage different levels of realism over different regions in the latent space. With extensive results in both photorealistic and non-photorealistic domains, we demonstrate qualitatively and quantitatively that our few-shot model automatically discovers correspondences between source and target domains and generates more diverse and realistic images than previous methods.

Cite

Text

Ojha et al. "Few-Shot Image Generation via Cross-Domain Correspondence." Conference on Computer Vision and Pattern Recognition, 2021. doi:10.1109/CVPR46437.2021.01060

Markdown

[Ojha et al. "Few-Shot Image Generation via Cross-Domain Correspondence." Conference on Computer Vision and Pattern Recognition, 2021.](https://mlanthology.org/cvpr/2021/ojha2021cvpr-fewshot/) doi:10.1109/CVPR46437.2021.01060

BibTeX

@inproceedings{ojha2021cvpr-fewshot,
  title     = {{Few-Shot Image Generation via Cross-Domain Correspondence}},
  author    = {Ojha, Utkarsh and Li, Yijun and Lu, Jingwan and Efros, Alexei A. and Lee, Yong Jae and Shechtman, Eli and Zhang, Richard},
  booktitle = {Conference on Computer Vision and Pattern Recognition},
  year      = {2021},
  pages     = {10743-10752},
  doi       = {10.1109/CVPR46437.2021.01060},
  url       = {https://mlanthology.org/cvpr/2021/ojha2021cvpr-fewshot/}
}