Recent studies have shown remarkable success in image-to-image translation for two domains. However, existing approaches have limited scalability and robustness in handling more than two domains, since different models should be built independently for every pair of image domains. To address this limitation, we propose StarGAN, a novel and scalable approach that can perform image-to-image translations for multiple domains using only a single model. Such a unified model architecture of StarGAN allows simultaneous training of multiple datasets with different domains within a single network. This leads to StarGAN's superior quality of translated images compared to existing models as well as the novel capability of flexibly translating an input image to any desired target domain. We empirically demonstrate the effectiveness of our approach on a facial attribute transfer and a facial expression synthesis tasks.
核心贡献 · Key contributions
提出 StarGAN，一种统一的 GAN 框架，使用单个生成器和判别器实现多域图像翻译。 Proposes StarGAN, a unified GAN framework for multi-domain image translation using a single generator and discriminator.
引入掩码向量方法，支持跨不同标签集的数据集联合训练。 Introduces a mask vector method to enable joint training across datasets with different label sets.
在多属性迁移中实现优于现有模型的图像质量和灵活性。 Achieves superior image quality and flexibility in multi-attribute transfer compared to existing models.
展示可扩展性，参数数量远少于 CycleGAN 等跨域模型。 Demonstrates scalability with significantly fewer parameters than cross-domain models like CycleGAN.
在人脸属性迁移和表情合成任务上提供定性和定量结果。 Provides both qualitative and quantitative results on facial attribute transfer and expression synthesis tasks.
首个成功跨不同数据集的多域图像翻译工作。 First successful multi-domain image translation across different datasets.
局限 · Limitations
评估仅限于人脸属性迁移和表情合成任务。 Evaluation limited to facial attribute transfer and expression synthesis tasks.
需要预定义域标签；测试时无法处理未见域。 Requires predefined domain labels; cannot handle unseen domains at test time.
若掩码向量指定错误，联合训练可能失败。 Joint training with mask vector may fail if mask vector is incorrectly specified.
训练时间约一天（单 GPU），计算成本可能较高。 Training time is about one day on a single GPU, which may be computationally expensive.
未探索非人脸域或更复杂场景的性能。 Performance on non-face domains or more complex scenes is not explored.
论文章节 · Sections（共 15）
摘要Abstract
1 引言1 Introduction
2 相关工作2 Related Work
3 Star 生成对抗网络3 Star Generative Adversarial Networks