StarGAN: Unified Generative Adversarial Networks for Multi-domain Image-to-Image Translation

By Yunjey Choi, Min-Je Choi, Munyoung Kim, Jung-Woo Ha, Sunghun Kim, J. Choo. 2017

3,914 citations recorded by the source corpus.

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.

Subfields: Computer Vision and Pattern Recognition, Generative Models

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Computer Vision and Pattern Recognition

In Generative Models

Citations in the corpus

CNN-Generated Images Are Surprisingly Easy to Spot… for Now (2019, methodology): “ProGAN [19] StyleGAN [20] BigGAN [7] CycleGAN [48] StarGAN [10] GauGAN [29] CRN [9] IMLE [23] SITD [8] Super-res.”
Large Scale Image Completion via Co-Modulated Generative Adversarial Networks (2021, background): “The unpaired setting is also investigated when paired training data is not available (Choi et al., 2018; Huang et al., 2018; Kim et al., 2019; Lazarow et al., 2017; Liu et al., 2017; Yi et al., 2017; Zhao et al., 2020b; Zhu et al., 2017a).”
Multimodal Unsupervised Image-to-Image Translation (2018, methodology): “Our problem has some connections with multi-domain image-to-image translation [19, 49, 50].”
Recent Progress on Generative Adversarial Networks (GANs): A Survey (2019, background): “[52] proposed a StarGAN, which can solve the problem of image translation among”
Piggyback GAN: Efficient Lifelong Learning for Image Conditioned Generation (2021, background): “However, for generative models, it is quite common that we want to translate the same image to different domains for different tasks [14,11,44,8].”
Palette: Image-to-Image Diffusion Models (2021, methodology): “, 2016], multi-domain translation [Choi et al., 2018], few shot translation [Liu et al.”
StyleFlow: Attribute-conditioned Exploration of StyleGAN-Generated Images using Conditional Continuous Normalizing Flows (2020, background): “In the context of faces, StarGAN (Choi et al. 2018, 2019) proposes a GAN architecture that considers face a ributes such as hair color, gender, and age.”
Encoding in Style: a StyleGAN Encoder for Image-to-Image Translation (2020, background): “Since then, their work has been extended for many scenarios: high-resolution synthesis [42], unsupervised learning [31, 50, 24, 29], multi-modal image synthesis [7, 15, 51], multi-domain image synthesis [6, 7], and conditional image synthesis [35, 53, 32, 5, 26].”
Few-Shot Unsupervised Image-to-Image Translation (2019, methodology): “As none of the prior unsupervised image-to-image translation methods are designed for the setting, we build a baseline by extending the StarGAN method [8], which is the state of the art for multi-class unsupervised imageto-image translation.”
FakeSpotter: A Simple Baseline for Spotting AI-Synthesized Fake Faces (2019, background): “Recent work in image synthesis like PGGAN [1], StyleGAN [2], and StarGAN [3] achieved remarkable results in face image synthesis.”
Joint Discriminative and Generative Learning for Person Re-Identification (2019, methodology): “StarGAN [6] is used in [56] to generate pedestrian images with different camera styles.”
Unpaired Image-to-Image Translation via Latent Energy Transport (2020, methodology): “We show the qualitative comparison between our method and StarGAN v2 in Figure 6 and Figure 7.”
Soft-Label Dataset Distillation and Text Dataset Distillation (2019, background): “They are primarily used to produce images that closely mimic those coming from the true distribution (Ledig et al., 2017; Goodfellow et al., 2014; Choi et al., 2018; Radford et al., 2015).”
STGAN: A Unified Selective Transfer Network for Arbitrary Image Attribute Editing (2019, methodology): “2, the target attribution vector attt is exactly the same as on the source one atts, but StarGAN [7] and AttGAN [11] may manipulate some unchanged attributes by mistake.”
GAN Prior Embedded Network for Blind Face Restoration in the Wild (2021, background): “Many following works, such as unsupervised learning [52], disentangled learning [26], few-shot learning [32], high resolution image synthesis [43], multi-domain translation [8], multimodal translation [53], have been proposed to extend cGAN to different scenarios.”