StarGAN v2: Diverse Image Synthesis for Multiple Domains

Abstract

A good image-to-image translation model should learn a mapping between different visual domains while satisfying the following properties: 1) diversity of generated images and 2) scalability over multiple domains. Existing methods address either of the issues, having limited diversity or multiple models for all domains. We propose StarGAN v2, a single framework that tackles both and shows significantly improved results over the baselines. Experiments on CelebA-HQ and a new animal faces dataset (AFHQ) validate our superiority in terms of visual quality, diversity, and scalability. To better assess image-to-image translation models, we release AFHQ, high-quality animal faces with large inter- and intra-domain differences. The code, pretrained models, and dataset can be found at https://github.com/clovaai/stargan-v2.

arXiv commentAccepted to CVPR 2020

References in corpus (20)

• Batch Normalization: Accelerating Deep Network Training by Reducing Internal Covariate Shift
• Conditional Generative Adversarial Nets
• GANs Trained by a Two Time-Scale Update Rule Converge to a Local Nash Equilibrium
• Large Scale GAN Training for High Fidelity Natural Image Synthesis
• Instance Normalization: The Missing Ingredient for Fast Stylization
• Generative Adversarial Text to Image Synthesis
• Conditional Image Synthesis With Auxiliary Classifier GANs
• Progressive Growing of GANs for Improved Quality, Stability, and Variation
• ESRGAN: Enhanced Super-Resolution Generative Adversarial Networks
• Which Training Methods for GANs do actually Converge?
• Learning to Discover Cross-Domain Relations with Generative Adversarial Networks
• Layer Normalization
• Large Scale Adversarial Representation Learning
• Augmented CycleGAN: Learning Many-to-Many Mappings from Unpaired Data
• DRIT++: Diverse Image-to-Image Translation via Disentangled Representations
• NSML: Meet the MLaaS platform with a real-world case study
• The Unusual Effectiveness of Averaging in GAN Training
• High-Fidelity Image Generation With Fewer Labels
• NSML: A Machine Learning Platform That Enables You to Focus on Your Models
• Multi-mapping Image-to-Image Translation via Learning Disentanglement

Cited by in corpus (18)

• Training Generative Adversarial Networks with Limited Data
• DAugNet: Unsupervised, Multi-source, Multi-target, and Life-long Domain Adaptation for Semantic Segmentation of Satellite Images
• I2V-GAN: Unpaired Infrared-to-Visible Video Translation
• Extracting quantitative biological information from brightfield cell images using deep learning
• StyleGAN2 Distillation for Feed-forward Image Manipulation
• K-Hairstyle: A Large-scale Korean Hairstyle Dataset for Virtual Hair Editing and Hairstyle Classification
• Frequency-Controlled Diffusion Model for Versatile Text-Guided Image-to-Image Translation
• Decentralized Attribution of Generative Models
• TuiGAN: Learning Versatile Image-to-Image Translation with Two Unpaired Images
• Murine AI excels at cats and cheese: Structural differences between human and mouse neurons and their implementation in generative AIs
• F2GAN: Fusing-and-Filling GAN for Few-shot Image Generation
• SLGAN: Style- and Latent-guided Generative Adversarial Network for Desirable Makeup Transfer and Removal
• Integrating Categorical Semantics into Unsupervised Domain Translation
• Few-shot Knowledge Transfer for Fine-grained Cartoon Face Generation
• Translate the Facial Regions You Like Using Region-Wise Normalization
• Synthesizing Photorealistic Images with Deep Generative Learning
• MCMI: Multi-Cycle Image Translation with Mutual Information Constraints
• Style-Restricted GAN: Multi-Modal Translation with Style Restriction Using Generative Adversarial Networks