Data to 14 Oct 2026

Seeing What a GAN Cannot Generate

arXiv 1910.11626 ↗First posted 24 Oct 2019cs.CV cs.GR cs.LGcited by 148

Not on the map. Why?

Cited by 32 in Diffusion models, but placed by its own abstract, which didn’t fit.

Abstract (arXiv)

Despite the success of Generative Adversarial Networks (GANs), mode collapse remains a serious issue during GAN training. To date, little work has focused on understanding and quantifying which modes have been dropped by a model. In this work, we visualize mode collapse at both the distribution level and the instance level. First, we deploy a semantic segmentation network to compare the distribution of segmented objects in the generated images with the target distribution in the training set. Differences in statistics reveal object classes that are omitted by a GAN. Second, given the identified omitted object classes, we visualize the GAN's omissions directly. In particular, we compare specific differences between individual photos and their approximate inversions by a GAN. To this end, we relax the problem of inversion and solve the tractable problem of inverting a GAN layer instead of the entire generator. Finally, we use this framework to analyze several recent GANs trained on multiple datasets and identify their typical failure cases.

Key facts. Cited by 148 papers in the map (data through 14 Oct 2026). First posted on arXiv 24 Oct 2019.

Builds on

No reference list yet.

Built on by

Cited by year (2026 to 14 Oct)

Citing papers per year: 2020: 1; 2021: 45; 2022: 29; 2023: 33; 2024: 18; 2025: 14; 2026: 8.

148 papers.

Sort:

By area:

2020Denoising Diffusion Implicit Modelscited by 9,781
2021SDEdit: Guided Image Synthesis and Editing with Stochastic Differential Equationscited by 1,937
2021GAN Inversion: A Surveycited by 367
2023RealFusion: 360deg Reconstruction of Any Object from a Single Imagecited by 306
2021EditGAN: High-Precision Semantic Image Editingcited by 156
2021Semantic Segmentation with Generative Models: Semi-Supervised Learning and Strong Out-of-Domain Generalizationcited by 141
2021Pros and Cons of GAN Evaluation Measures: New Developmentscited by 123
2022Mildly Conservative Q-Learning for Offline Reinforcement Learningcited by 112
2021HIVE: Evaluating the Human Interpretability of Visual Explanationscited by 105
2021Gradient Inversion with Generative Image Priorcited by 101
2021The Majority Can Help The Minority: Context-rich Minority Oversampling for Long-tailed Classificationcited by 100
2021D2C: Diffusion-Denoising Models for Few-shot Conditional Generationcited by 96

38 more

2021Learning by Watching: Physical Imitation of Manipulation Skills from Human Videoscited by 87
2021HyperInverter: Improving StyleGAN Inversion via Hypernetworkcited by 85
2021CLIP2StyleGAN: Unsupervised Extraction of StyleGAN Edit Directionscited by 78
2021Defect-GAN: High-Fidelity Defect Synthesis for Automated Defect Inspectioncited by 77
2021FuseDream: Training-Free Text-to-Image Generation with Improved CLIP+GAN Space Optimizationcited by 71
2021Intermediate Layer Optimization for Inverse Problems using Deep Generative Modelscited by 70
2021Unsupervised 3D Shape Completion through GAN Inversioncited by 70
2021Mind the Gap: Domain Gap Control for Single Shot Domain Adaptation for Generative Adversarial Networkscited by 68
2023CoLa-Diff: Conditional Latent Diffusion Model for Multi-Modal MRI Synthesiscited by 65
2022"Help Me Help the AI": Understanding How Explainability Can Support Human-AI Interactioncited by 63
2021GAN-Supervised Dense Visual Alignmentcited by 62
2024FakeInversion: Learning to Detect Images from Unseen Text-to-Image Models by Inverting Stable Diffusioncited by 61
2023Latent diffusion models for generative precipitation nowcasting with accurate uncertainty quantificationcited by 56
2021Using latent space regression to analyze and leverage compositionality in GANscited by 55
2022Generated Faces in the Wild: Quantitative Comparison of Stable Diffusion, Midjourney and DALL-E 2cited by 55
2021Labels4Free: Unsupervised Segmentation using StyleGANcited by 53
2021Ensembling with Deep Generative Viewscited by 53
2021Finding an Unsupervised Image Segmenter in Each of Your Deep Generative Modelscited by 44
20223D GAN Inversion with Pose Optimizationcited by 44
2023Make-A-Volume: Leveraging Latent Diffusion Models for Cross-Modality 3D Brain MRI Synthesiscited by 43
2022Spatially-Adaptive Multilayer Selection for GAN Inversion and Editingcited by 39
2024Aligned Datasets Improve Detection of Latent Diffusion-Generated Imagescited by 36
2021MarioNette: Self-Supervised Sprite Learningcited by 35
2022BlobGAN: Spatially Disentangled Scene Representationscited by 35
2021E2Style: Improve the Efficiency and Effectiveness of StyleGAN Inversioncited by 33
2022Studying Bias in GANs through the Lens of Racecited by 31
2021Unsupervised Object Learning via Common Fatecited by 29
2023Bayesian Conditional Diffusion Models for Versatile Spatiotemporal Turbulence Generationcited by 29
2024SwiftEdit: Lightning Fast Text-Guided Image Editing via One-Step Diffusioncited by 28
2021Cycle-Consistent Inverse GAN for Text-to-Image Synthesiscited by 27
2022Generative Visual Prompt: Unifying Distributional Control of Pre-Trained Generative Modelscited by 27
2023Accurate Fine-Grained Segmentation of Human Anatomy in Radiographs via Volumetric Pseudo-Labelingcited by 26
2021IMAGINE: Image Synthesis by Image-Guided Model Inversioncited by 25
2023Boundary Guided Learning-Free Semantic Control with Diffusion Modelscited by 23
2023Precipitation nowcasting with generative diffusion modelscited by 23
2021LTT-GAN: Looking Through Turbulence by Inverting GANscited by 21
2022Polarity Sampling: Quality and Diversity Control of Pre-Trained Generative Networks via Singular Valuescited by 21
2023Generative AI for End-to-End Limit Order Book Modelling: A Token-Level Autoregressive Generative Model of Message Flow Using a Deep State Space Networkcited by 20

50 most cited of 148 All 148 (JSON)