Domain adaptation of GANs is a problem of fine-tuning the state-of-the-art GAN models (e.g. StyleGAN) pretrained on a large dataset to a specific domain with few samples (e.g. painting faces, sketches, etc.). While there are a great number of methods that tackle this problem in different ways, there are still many important questions that remain unanswered. In this paper, we provide a systematic and in-depth analysis of the domain adaptation problem of GANs, focusing on the StyleGAN model. First, we perform a detailed exploration of the most important parts of StyleGAN that are responsible for adapting the generator to a new domain depending on the similarity between the source and target domains. As a result of this in-depth study, we propose new efficient and lightweight parameterizations of StyleGAN for domain adaptation. Particularly, we show there exist directions in StyleSpace (StyleDomain directions) that are sufficient for adapting to similar domains and they can be reduced further. For dissimilar domains, we propose Affine$+$ and AffineLight$+$ parameterizations that allows us to outperform existing baselines in few-shot adaptation with low data regime. Finally, we examine StyleDomain directions and discover their many surprising properties that we apply for domain mixing and cross-domain image morphing.

翻译：摘要：生成对抗网络（GAN）的领域自适应问题，是指在拥有少量样本的特定领域（例如人物画像、素描等）中，微调在大规模数据集上预训练的先进GAN模型（如StyleGAN）的过程。尽管已有大量方法以不同方式解决该问题，但仍有许多重要问题悬而未决。本文对GAN的领域自适应问题进行了系统而深入的分析，重点关注StyleGAN模型。首先，我们详细探究了StyleGAN中负责根据源域与目标域相似性将生成器适应新领域的最关键部分。基于这项深入研究，我们提出了适用于领域自适应的新型高效轻量化StyleGAN参数化方法。具体而言，我们发现StyleSpace中存在足以适应相似领域的特定方向（即StyleDomain方向），且这些方向可进一步精简。针对不相似领域，我们提出Affine+和AffineLight+参数化方法，能够在低数据量的少样本自适应场景中超越现有基线。最后，我们考察了StyleDomain方向，发现了诸多令人惊奇的特性，并将其应用于领域混合与跨域图像变形。