cGANs with Projection Discriminator

Abstract

We propose a novel, projection based way to incorporate the conditional information into the discriminator of GANs that respects the role of the conditional information in the underlining probabilistic model. This approach is in contrast with most frameworks of conditional GANs used in application today, which use the conditional information by concatenating the (embedded) conditional vector to the feature vectors. With this modification, we were able to significantly improve the quality of the class conditional image generation on ILSVRC2012 (ImageNet) 1000-class image dataset from the current state-of-the-art result, and we achieved this with a single pair of a discriminator and a generator. We were also able to extend the application to super-resolution and succeeded in producing highly discriminative super-resolution images. This new structure also enabled high quality category transformation based on parametric functional transformation of conditional batch normalization layers in the generator.

arXiv commentPublished as a conference paper at ICLR 2018

References in corpus (12)

• Conditional Generative Adversarial Nets
• GANs Trained by a Two Time-Scale Update Rule Converge to a Local Nash Equilibrium
• Spectral Normalization for Generative Adversarial Networks
• Generative Adversarial Text to Image Synthesis
• Conditional Image Synthesis With Auxiliary Classifier GANs
• Deep Generative Image Models using a Laplacian Pyramid of Adversarial Networks
• Improved Training of Wasserstein GANs
• Improved Techniques for Training GANs
• Adversarially Learned Inference
• Invertible Conditional GANs for image editing
• Hierarchical Implicit Models and Likelihood-Free Variational Inference
• Semi-supervised Conditional GANs

Cited by in corpus (90)

• Diffusion Models Beat GANs on Image Synthesis
• Generative Adversarial Network in Medical Imaging: A Review
• Training Generative Adversarial Networks with Limited Data
• Perceptual Adversarial Networks for Image-to-Image Transformation
• How Generative Adversarial Networks and Their Variants Work: An Overview
• Self-supervised learning methods and applications in medical imaging analysis: A survey
• A Review on Generative Adversarial Networks: Algorithms, Theory, and Applications
• Adversarial Video Generation on Complex Datasets
• Classification Accuracy Score for Conditional Generative Models
• Video-to-Video Synthesis
• Train Sparsely, Generate Densely: Memory-efficient Unsupervised Training of High-resolution Temporal GAN
• InSituNet: Deep Image Synthesis for Parameter Space Exploration of Ensemble Simulations
• Weakly Supervised Disentanglement with Guarantees
• SESAME: Semantic Editing of Scenes by Adding, Manipulating or Erasing Objects
• Memory Replay GANs: learning to generate images from new categories without forgetting
• A Systematic Survey of Regularization and Normalization in GANs
• GACELA -- A generative adversarial context encoder for long audio inpainting
• Is Generator Conditioning Causally Related to GAN Performance?
• Generative Artificial Intelligence Meets Synthetic Aperture Radar: A Survey
• PConv: Simple yet Effective Convolutional Layer for Generative Adversarial Network
• Simple yet Effective Way for Improving the Performance of GAN
• Landmark Assisted CycleGAN for Cartoon Face Generation
• DeepFlow: History Matching in the Space of Deep Generative Models
• Ultra-High-Resolution Detector Simulation with Intra-Event Aware GAN and Self-Supervised Relational Reasoning
• Generative Adversarial Networks: A Survey Towards Private and Secure Applications
• MarioNETte: Few-shot Face Reenactment Preserving Identity of Unseen Targets
• TF-Replicator: Distributed Machine Learning for Researchers
• A unified framework for 21cm tomography sample generation and parameter inference with Progressively Growing GANs
• Transferring GANs: generating images from limited data
• Towards Adversarial Denoising of Radar Micro-Doppler Signatures
• What You See is Not What the Network Infers: Detecting Adversarial Examples Based on Semantic Contradiction
• Mimicry: Towards the Reproducibility of GAN Research
• Exploring DeshuffleGANs in Self-Supervised Generative Adversarial Networks
• PcDGAN: A Continuous Conditional Diverse Generative Adversarial Network For Inverse Design
• GRDN:Grouped Residual Dense Network for Real Image Denoising and GAN-based Real-world Noise Modeling
• Anisotropic Stroke Control for Multiple Artists Style Transfer
• UGAN: Untraceable GAN for Multi-Domain Face Translation
• Generating unrepresented proportions of geological facies using Generative Adversarial Networks
• Training Generative Adversarial Networks by Solving Ordinary Differential Equations
• Generative Convolution Layer for Image Generation
• LatteGAN: Visually Guided Language Attention for Multi-Turn Text-Conditioned Image Manipulation
• Unbiased Auxiliary Classifier GANs with MINE
• Generative Feature Replay For Class-Incremental Learning
• Task-Aware Feature Generation for Zero-Shot Compositional Learning
• Towards Efficiently Evaluating the Robustness of Deep Neural Networks in IoT Systems: A GAN-based Method
• Triple Generative Adversarial Networks
• Learning Spatial Pyramid Attentive Pooling in Image Synthesis and Image-to-Image Translation
• Few-shot Video-to-Video Synthesis
• Group Equivariant Generative Adversarial Networks
• MineGAN: effective knowledge transfer from GANs to target domains with few images
• JGAN: A Joint Formulation of GAN for Synthesizing Images and Labels
• Image Synthesis From Reconfigurable Layout and Style
• Causal Adversarial Network for Learning Conditional and Interventional Distributions
• Generative Adversarial Zero-Shot Relational Learning for Knowledge Graphs
• StarGAN-ZSVC: Towards Zero-Shot Voice Conversion in Low-Resource Contexts
• Site-Specific Ground Motion Generative Model for Crustal Earthquakes in Japan Based on Generative Adversarial Networks
• SocialInteractionGAN: Multi-person Interaction Sequence Generation
• cGANs with Conditional Convolution Layer
• Attentive Normalization for Conditional Image Generation
• Scalable Balanced Training of Conditional Generative Adversarial Neural Networks on Image Data
• Continuous Conditional Generative Adversarial Networks (cGAN) with Generator Regularization
• Taking Control of Intra-class Variation in Conditional GANs Under Weak Supervision
• Towards Audio to Scene Image Synthesis using Generative Adversarial Network
• F2GAN: Fusing-and-Filling GAN for Few-shot Image Generation
• IntersectGAN: Learning Domain Intersection for Generating Images with Multiple Attributes
• Sketch-to-Art: Synthesizing Stylized Art Images From Sketches
• Towards Amortized Ranking-Critical Training for Collaborative Filtering
• On Computation and Generalization of GANs with Spectrum Control
• Adversarial Feature Desensitization
• Set Distribution Networks: a Generative Model for Sets of Images
• Learning View Priors for Single-view 3D Reconstruction
• Interactive Text2Pickup Network for Natural Language based Human-Robot Collaboration
• On The Distribution of Penultimate Activations of Classification Networks
• Data Augmentation via Mixed Class Interpolation using Cycle-Consistent Generative Adversarial Networks Applied to Cross-Domain Imagery
• Intrinsic Autoencoders for Joint Neural Rendering and Intrinsic Image Decomposition
• Novel View Synthesis on Unpaired Data by Conditional Deformable Variational Auto-Encoder
• Disentangling the Spatial Structure and Style in Conditional VAE
• MixDefense: A Defense-in-Depth Framework for Adversarial Example Detection Based on Statistical and Semantic Analysis
• Detecting Invasive Ductal Carcinoma with Semi-Supervised Conditional GANs
• Enhancing Stability in Training Conditional Generative Adversarial Networks via Selective Data Matching
• Venn GAN: Discovering Commonalities and Particularities of Multiple Distributions
• Adversarial Learning of Intermediate Acoustic Feature for End-to-End Lightweight Text-to-Speech
• Disentangling Latent Space for VAE by Label Relevant/Irrelevant Dimensions
• Perceiving Music Quality with GANs
• Dual Projection Generative Adversarial Networks for Conditional Image Generation
• Generative Adversarial Learning via Kernel Density Discrimination
• MixPoet: Diverse Poetry Generation via Learning Controllable Mixed Latent Space
• Identifying and Exploiting Structures for Reliable Deep Learning
• Multiclass non-Adversarial Image Synthesis, with Application to Classification from Very Small Sample
• Latent Transformations for Object View Points Synthesis