Research map / Image generation / GANs
GAN-based image manipulation
Apply GANs to editing, translation, and domain-specific image synthesis.
90 accepted papers at ICML, NeurIPS, ICLR, CVPR and AAAI (2017–2026) · an approach in GANs, keyword cluster “synthesis · latent · face”
State History Tried Checklist Preprints Who Export
Where it stands
Not enough benchmark data Fewer than four papers report the same benchmark, so there is no record to follow.
10 papers in 2023–24 and 7 in 2025–26 (at ICML, NeurIPS, ICLR, CVPR and AAAI); its share of the topic grew, from 17.2% to 28.0%.
How this is decided
A benchmark is tracked when at least four papers of this approach report their own method on the same dataset and metric in at least two different years. Its record is the best number reported up to 2024 compared with the best reported in 2025–26: still improving means at least 1% better, flat means at most 0.2% better with at least two papers from 2025–26 trying. Numbers on a different scale from the rest (a fraction among percentages) and far outliers are left out. Results come from different papers with different setups, so they are signals, not head-to-head comparisons. The share of papers is counted at ICML, NeurIPS, ICLR, CVPR and AAAI.
How it got here
Most cited paper of each year
- 2017 · Learning From Simulated and Unsupervised Images Through Adversarial Training (CVPR 2017 · 1,971 citations)
- 2018 · Image Blind Denoising With Generative Adversarial Network Based Noise Modeling (CVPR 2018 · 637 citations)
- 2019 · A Style-Based Generator Architecture for Generative Adversarial Networks (CVPR 2019 · 11,298 citations)
- 2020 · Watch Your Up-Convolution: CNN Based Generative Deep Neural Networks Are Failing to Reproduce Spectral Distributions (CVPR 2020 · 456 citations)
- 2021 · Adversarial Generation of Continuous Images (CVPR 2021 · 124 citations)
- 2022 · Think Twice Before Detecting GAN-Generated Fake Images From Their Spectral Domain Imprints (CVPR 2022 · 69 citations)
- 2023 · Generative Image Inpainting with Segmentation Confusion Adversarial Training and Contrastive Learning (AAAI 2023 · 44 citations)
- 2024 · Virtual Immunohistochemistry Staining for Histological Images Assisted by Weakly-supervised Learning (CVPR 2024 · 20 citations)
- 2025 · AvatarArtist: Open-Domain 4D Avatarization (CVPR 2025 · 4 citations)
- 2026 · New Synthetic Goldmine: Hand Joint Angle-Driven EMG Data Generation Framework for Micro-Gesture Recognition (AAAI 2026 · 1 citation)
Earliest papers here
- Learning From Simulated and Unsupervised Images Through Adversarial Training (CVPR 2017 · 1,971 citations)
- Age Progression/Regression by Conditional Adversarial Autoencoder (CVPR 2017 · 1,124 citations)
- Disentangled Representation Learning GAN for Pose-Invariant Face Recognition (CVPR 2017 · 1,090 citations)
What has been tried
84 papers here report 775 measured changes (a component added, removed or swapped, compared with a baseline): helped 75%, no clear effect 5%, hurt 16%, mixed 4%. Papers mostly publish changes that helped, so "helped" is over-represented.
Ablations in recent papers
| What was changed | Measured on | Result | Effect | Paper |
|---|---|---|---|---|
| add PPIE to confusion-GAN | HCI · FID | 105.4 -> 103.3 | Helped | Virtual Immunohistochemistry Staining for Histological Images Assisted by Weakly-supervised Learning (CVPR 2024 · 20 citations) |
| Use iterative generation instead of one training iteration | Places365-Standard 256×256 · FID | 2.95 -> 2.36 | Helped | Image Inpainting via Iteratively Decoupled Probabilistic Modeling (ICLR 2024 · 5 citations) |
| replace the Next3D CNN renderer with the motion-aware cross-domain renderer | FFHQ cross-reenactment · FID | 130.72 -> 68.69 | Helped | AvatarArtist: Open-Domain 4D Avatarization (CVPR 2025 · 4 citations) |
| Add low-frequency loss L_lf | CycleGAN · image FID | 60.62 -> 59.85 | Helped | Spectrum Translation for Refinement of Image Generation (STIG) Based on Contrastive Learning and Spectral Filter Profile (AAAI 2024 · 4 citations) |
| Increase sampling steps from 1 to 10 | ZeroEGGs · FGD | 85.44 -> 12.57 | Helped | DIDiffGes: Decoupled Semi-Implicit Diffusion Models for Real-time Gesture Generation from Speech (AAAI 2025 · 3 citations) |
| remove prediction matching | DBP15K ZH-EN · Hits@1 | .761 -> .045 | Hurt | Revisit and Outstrip Entity Alignment: A Perspective of Generative Models (ICLR 2024 · 3 citations) |
| Add second-stage inpainting instead of directly pasting rectified parts | Human-part integration · visual integration quality | Helped | RealisHuman: A Two-Stage Approach for Refining Malformed Human Parts in Generated Images (AAAI 2025 · 2 citations) | |
| add training-free alignment (TAA: transformation-averaged gradients) during inversion | High-resolution CelebA and FaceScrub · Acc@1 | +5.24 to +12.71 across models | Helped | Generative Model Inversion Through the Lens of the Manifold Hypothesis (NeurIPS 2025 · 0 citations) |
What reviewers will expect
Datasets and baselines named in the 5 papers of this approach from 2025–2026 that have a full-text summary. Leaving out the common ones is a frequent review question.
Datasets
- FFHQ · 2 of 5 papers
Recent preprints
Loading recent arXiv preprints…
Who publishes on it
Institutions on the most papers of this approach since 2023 (of 17 papers; from OpenAlex and the papers' affiliations).
- Huawei · 3 papers
- Zhejiang University · 2 papers
- Mohamed bin Zayed University of Artificial Intelligence · 2 papers
- University of Melbourne · 2 papers
- Alibaba Group · 2 papers
- Hong Kong University of Science and Technology · 2 papers
Use it
Download BibTeX of 32 key papersOpen the topic in the interactive map
Get alerts. Describe your idea in scime atlas and save it (free scime account): you get a weekly email when new preprints land close to it.
Related approaches
Same keyword cluster
- Large-scale GAN latent analysis (108 papers)
- Conditional GANs (25 papers)
- GAN vocoders and audio synthesis (13 papers)
Elsewhere in the topic
- Stabilized GAN training (144 papers)
- Generative adversarial network theory (62 papers)
- GAN latent optimization and evaluation (52 papers)
- GANs for text generation (21 papers)
Benchmark numbers, ablations and baseline numbers are extracted by AI from each paper's full text and can be wrong; the paper is the authority. Data: results index of 2026-11-02, citation counts from OpenAlex (2026-10-09).