Computer Visions

---

🖼 Generative Adversarial Networks : Paper Review (Github)

GAN Basics

• GAN: Generative Adversarial Networks (NIPS 2014) : arxiv
• DCGAN: Unsupervised Representation Learning with Deep Convolutional Generative Adversarial Networks (ICLR 2016) : arxiv

Conditional GAN

• CGAN: Conditional Generative Adversarial Nets (2014) : arxiv

• ACGAN: Conditional Image Synthesis With Auxiliary Classifier GANs (ICML 2017) : arxiv

• Supervised Approach

  • Pix2Pix: Image-to-Image Translation with Conditional Adversarial Networks (CVPR 2017) : arxiv

  • GAN Dissection: Visualizing and Understanding Generative Adversarial Networks (ICLR 2019) : arxiv

  • SPADE: Semantic Image Synthesis with Spatially Adaptive Normalization (CVPR 2019) : arxiv

• Unsupervised Approach

  • CycleGAN: Unpaired Image-to-Image Translation using Cycle-Consistent Adversarial Networks (ICCV 2017) : arxiv

  • FUNIT: Few-Shot Unsupervised Image-to-Image Translation (ICCV 2019) : arxiv

  • COCO-FUNIT: Few-Shot Unsupervised Image Translation with a Content Conditioned Style Encoder (ECCV 2020) : arxiv
  • HiGAN: Semantic Hierarchy Emerges in Deep Generative Representations for Scene Synthesis (IJCV 2020) : arxiv, project page
• Multi Domain

  • BicycleGAN: Toward Multimodal Image-to-Image Translation (NIPS 2017) : arxiv

  • StarGAN: Unified Generative Adversarial Networks for Multi-Domain Image-to-Image Translation (CVPR 2018) : arxiv

    • StarGAN v2: Diversity Image Synthesis for Multiple Domains (CVPR 2020) : arxiv

  • MUNIT : Multi-Modal Unsupervised Image-to-Image Translation (ECCV 2018) : arxiv

GAN Architecture

• PGGAN: Progressive Growing of GANs for Improved Quality, Stability, and Variation (ICLR 2018) : arxiv

• StyleGAN: A Style-Based Generator Architecture for Generative Adversarial Networks (CVPR 2019) : arxiv

  • StyleGAN v2: Analyzing and Improving the Image Quality of StyleGAN (CVPR 2020) : arxiv

  • StyleGAN-ADA: Training Generative Adversarial Networks with Limited Data (NeurlPS 2020) : arxiv

  • StyleGAN v3: Alias-Free Generative Adversarial Networks (NeurIPS 2021) : (arxiv, code, project)

• BigGAN: Large Scale GAN Training for High Fidelity Natural Image Synthesis (2019) : arxiv, code

Text-to-Image

• Generative Adversarial Text to Image Synthesis (ICML 2016) : arxiv

• TediGAN: Text-Guided Diverse Face Image Generation and Manipulation (CVPR 2021) : arxiv

• StyleCLIP: Text-Driven Manipulation of StyleGAN Imagery (arXiv 2021) : arxiv

• DALLE: Zero-Shot Text-to-Image Generation (ICML 2021) : arxiv, project page
• Paint by Word (2021) : arxiv

Improved Training Techniques

• SS-GAN: Self-Supervised GANs via Auxiliary Rotation Loss (CVPR 2019) : paper

• CR-GAN: Consistency Regularization for Generative Adversarial Networks (ICLR 2020) : arxiv

• ICR-GAN: Improved Consistency Regularization for GANs (AAAI 2021) : arxiv

GAN Inversion

1. Latent Optimization

   • Image2stylegan: How to embed images into the stylegan latent space? (ICCV 2019) : arxiv

   • Image2stylegan++: How to edit the embedded images? (CVPR 2020) : arxiv

   • StyleFlow: Attribute-conditioned Exploration of StyleGAN-Generated Images using Conditional Continuous Normalizing Flows (ACM TOG 2021) : arxiv, project page

   • BDInvert: GAN Inversion for Out-of-Range Images with Geometric Transformations (ICCV 2021) : arxiv, code

2. Encoder

   • ALAE: Adversarial latent autoencoders (CVPR 2020) : arxiv, code

   • pSp: Encoding in Style: a StyleGAN Encoder for Image-to-Image Translation (CVPR 2021) : arxiv

3. Hybrid approach

   • stylegan-encoder : code

   • IdInvert : In-Domain GAN Inversion for Real Image Editing (ECCV 2020) : arxiv

Disentangled Manipulation

• GANSpace: Discovering Interpretable GAN Controls (NeurIPS 2020) : arxiv, code

• GAN-Latent-Discovery: Unsupervised Discovery of Interpretable Directions in the GAN Latent Space (2020) : arxiv, code

• Editing in style: Uncovering the Local Semantics of GANs (CVPR 2020) : arxiv, code

• HiGAN: Semantic Hierarchy Emerges in Deep Generative Representations for Scene Synthesis (IJCV 2020) : arxiv, project page

• InterFaceGAN: Interpreting the Latent Space of GANs for Semantic Face Editing (CVPR 2020) : arxiv, project page

• CD3D: Cross-Domain and Disentangled Face Manipulation with 3D Guidance (2021) : arxiv, code

• GHFeat: Generative Hierarchical Features from Synthesizing Images (CVPR 2021) : arxiv, project page

• StyleSpace: StyleSpace Analysis: Disentangled Controls for StyleGAN Image Generation (2021) : arxiv, code

• StyleFlow: Attribute-conditioned Exploration of StyleGAN-Generated Images using Conditional Continuous Normalizing Flows (ACM TOG 2021) : arxiv, project page

• Hessian Penalty: A weak prior for unsupervised disentanglement (ECCV 2020) : arxiv, project page

• StyleGAN of All Trades: Image Manipulation with Only Pretrained StyleGAN (arxiv 2021) : arxiv, code

Image Editing

• StyleCLIP: Text-Driven Manipulation of StyleGAN Imagery (arXiv 2021) : arxiv, code

• sefa: Closed-Form Factorization of Latent Semantics in GANs (CVPR 2021) : arxiv

• EigenGAN: Layer-Wise Eigen-Learning for GANs : arxiv, code

• StyleMapGAN: Exploiting Spatial Dimensions of Latent in GAN for Real-time Image Editing (CVPR 2021) : arxiv, code

• SEAN: Image Synthesis with Semantic Region-Adaptive Normalization (CVPR 2020) : arxiv, code

• CDDFM3D: Cross-Domain and Disentangled Face Manipulation with 3D Guidance (2021) : arxiv, code, project

• MocoGAN-HD: A Good Image Generator Is What You Need for High-Resolution Video Synthesis (ICLR 2021) : arxiv, code, project

• StyleGAN-NADA: CLIP-Guided Domain Adaptation of Image Generators (arxiv 2021) : arxiv, project, code

• HyperStyle: StyleGAN Inversion with HyperNetworks for Real Image Editing (arxiv 2021) : arxiv, code

Webtoon/Anime GAN & Image Blending

• Cartoon-StylaGAN: Fine-tuning StyleGAN2 for Cartoon Face Generation (arxiv 2021) : arxiv, review, code
• BlendGAN: Implicitly GAN Blending for Arbitrary Stylized Face Generation (NeurIPS 2021) : arxiv, project, code
• HifiFace: 3D Shape and Semantic Prior Guided High Fidelity Face Swapping (IJCAI 2021) : arxiv
• AnimeGAN: A Novel Lightweight GAN for Photo Animation (ISICA 2019)
  • AnimeGANv2 (2021) : project, code
• StyleGAN of All Trades: Image Manipulation with Only Pretrained StyleGAN (arxiv 2021) : arxiv, code

Super Resolution

• BSRGAN: Designing a Practical Degradation Model for Deep Blind Image Super-Resolution (ICCV 2021) : arxiv, code

• Real-ESRGAN: Training Real-World Blind Super-Resolution with Pure Synthetic Data (ICCVW 2021): arxiv, code

3D GAN & Rendering

• HoloGAN: Unsupervised learning of 3D representations from natural images (ICCV 2019): paper, code

• CDDFM3D: Cross-Domain and Disentangled Face Manipulation with 3D Guidance (2021) : arxiv, code, project

• SofGAN: A Portrait Image Generator with Dynamic Styling (arxiv 2021): arxiv, project, code

• pi-GAN: Periodic Implicit Generative Adversarial Networks for 3D-Aware Image Synthesis (CVPR 2021): paper, project, code

• StyleGANRender: Image GANs meet Differentiable Rendering for Inverse Graphics and Interpretable 3D Neural Rendering (ICLR 2021) : arxiv, project page

• StyleNeRF: A Style-based 3D Aware Generator for High-resolution Image Synthesis (ICLR 2022): paper

  • NeRF: Representing Scenes as Neural Radiance Fields for View Synthesis (ECCV 2020) : arxiv, project

• CIPS-3D: A 3D-Aware Generator of GANs Based on Conditionally-Independent Pixel Synthesis (arxiv 2021): arxiv, code

• PIRenderer: Controllable Portrait Image Generation via Semantic Neural Rendering (ICCV 2021) : arxiv, project, code

---

Reference

• https://github.com/eriklindernoren/PyTorch-GAN