mlsanityData to 7 Oct 2026

Make It So: Steering StyleGAN for Any Image Inversion and Editing

arXiv 2304.14403 ↗First posted 27 Apr 2023cs.CV cs.GR cs.LGcited by 9

Not on the map. Why?

Abstract (arXiv)

StyleGAN's disentangled style representation enables powerful image editing by manipulating the latent variables, but accurately mapping real-world images to their latent variables (GAN inversion) remains a challenge. Existing GAN inversion methods struggle to maintain editing directions and produce realistic results. To address these limitations, we propose Make It So, a novel GAN inversion method that operates in the Z (noise) space rather than the typical W (latent style) space. Make It So preserves editing capabilities, even for out-of-domain images. This is a crucial property that was overlooked in prior methods. Our quantitative evaluations demonstrate that Make It So outperforms the state-of-the-art method PTI~citeroich2021pivotal by a factor of five in inversion accuracy and achieves ten times better edit quality for complex indoor scenes.

Key facts. Cited by 9 papers in the map (data through 7 Oct 2026). Builds on 50 of its 57 references that are in the map. First posted on arXiv 27 Apr 2023.

Builds on

50 of 57 references.

References · 50

  1. 2018The Unreasonable Effectiveness of Deep Features as a Perceptual Metriccited by 11,029
  2. 2018A Style-Based Generator Architecture for Generative Adversarial Networkscited by 5,986
  3. 2019Analyzing and Improving the Image Quality of StyleGANcited by 3,580
  4. 2016Learning without Forgettingcited by 2,465
  5. 2015LSUN: Construction of a Large-scale Image Dataset using Deep Learning with Humans in the Loopcited by 1,350
  6. 2019StarGAN v2: Diverse Image Synthesis for Multiple Domainscited by 1,095
  7. 2021StyleCLIP: Text-Driven Manipulation of StyleGAN Imagerycited by 990
  8. 2020Encoding in Style: a StyleGAN Encoder for Image-to-Image Translationcited by 727
  9. 2019Interpreting the Latent Space of GANs for Semantic Face Editingcited by 691
  10. 2019Image2StyleGAN: How to Embed Images Into the StyleGAN Latent Space?cited by 675
  11. 2020GANSpace: Discovering Interpretable GAN Controlscited by 612
  12. 2021Designing an Encoder for StyleGAN Image Manipulationcited by 542
38 more
  1. 2021Pivotal Tuning for Latent-based Editing of Real Imagescited by 427
  2. 2020In-Domain GAN Inversion for Real Image Editingcited by 420
  3. 2015Autoencoding beyond pixels using a learned similarity metriccited by 407
  4. 2020InterFaceGAN: Interpreting the Disentangled Face Representation Learned by GANscited by 393
  5. 2019Image2StyleGAN++: How to Edit the Embedded Images?cited by 356
  6. 2016Generative Visual Manipulation on the Natural Image Manifoldcited by 348
  7. 2020StyleSpace Analysis: Disentangled Controls for StyleGAN Image Generationcited by 331
  8. 2020Unsupervised Discovery of Interpretable Directions in the GAN Latent Spacecited by 288
  9. 2021ReStyle: A Residual-Based StyleGAN Encoder via Iterative Refinementcited by 248
  10. 2019On the "steerability" of generative adversarial networkscited by 247
  11. 2020Exploiting Deep Generative Prior for Versatile Image Restoration and Manipulationcited by 225
  12. 2020Swapping Autoencoder for Deep Image Manipulationcited by 215
  13. 2021HyperStyle: StyleGAN Inversion with HyperNetworks for Real Image Editingcited by 208
  14. 2021High-Fidelity GAN Inversion for Image Attribute Editingcited by 197
  15. 2019Image Processing Using Multi-Code GAN Priorcited by 185
  16. 2020Editing in Style: Uncovering the Local Semantics of GANscited by 183
  17. 2020Semantic Photo Manipulation with a Generative Image Priorcited by 178
  18. 2020Adversarial Latent Autoencoderscited by 155
  19. 2019Seeing What a GAN Cannot Generatecited by 148
  20. 2016Inverting The Generator Of A Generative Adversarial Networkcited by 140
  21. 2016Invertible Conditional GANs for image editingcited by 118
  22. 2016Neural Photo Editing with Introspective Adversarial Networkscited by 109
  23. 2021Exploiting Spatial Dimensions of Latent in GAN for Real-time Image Editingcited by 91
  24. 2020Generative Hierarchical Features from Synthesizing Imagescited by 83
  25. 2020Transforming and Projecting Images into Class-conditional Generative Networkscited by 76
  26. 2021Using latent space regression to analyze and leverage compositionality in GANscited by 55
  27. 2021GAN Inversion for Out-of-Range Images with Geometric Transformationscited by 55
  28. 2021Ensembling with Deep Generative Viewscited by 53
  29. 2021WarpedGANSpace: Finding non-linear RBF paths in GAN latent spacecited by 48
  30. 2022Spatially-Adaptive Multilayer Selection for GAN Inversion and Editingcited by 39
  31. 2021Dual Contrastive Loss and Attention for GANscited by 30
  32. 2021StyleFusion: A Generative Model for Disentangling Spatial Segmentscited by 28
  33. 2021From Continuity to Editability: Inverting GANs with Consecutive Imagescited by 25
  34. 2022High-fidelity GAN Inversion with Padding Spacecited by 23
  35. 2022Editing Out-of-domain GAN Inversion via Differential Activationscited by 14
  36. 2022IntereStyle: Encoding an Interest Region for Robust StyleGAN Inversioncited by 8
  37. 2022Cycle Encoding of a StyleGAN Encoder for Improved Reconstruction and Editabilitycited by 7
  38. 2022Overparameterization Improves StyleGAN Inversioncited by 4

Built on by

Cited by year

Citing papers per year: 2023: 2; 2024: 6; 2025: 1.

9 papers.

  1. 2023Generative Models: What Do They Know? Do They Know Things? Let's Find Out!cited by 50
  2. 2023StyleGAN knows Normal, Depth, Albedo, and Morecited by 44
  3. 2024LumiNet: Latent Intrinsics Meets Diffusion Models for Indoor Scene Relightingcited by 37
  4. 2024Latent Intrinsics Emerge from Training to Relightcited by 17
  5. 2024ZeroComp: Zero-shot Object Compositing from Image Intrinsics via Diffusioncited by 16
  6. 2024Retinex-Diffusion: On Controlling Illumination Conditions in Diffusion Models via Retinex Theorycited by 7
  7. 2024MyTimeMachine: Personalized Facial Age Transformationcited by 5
  8. 2025AgeBooth: Controllable Facial Aging and Rejuvenation via Diffusion Modelscited by 3
  9. 2024The Devil is in the Details: StyleFeatureEditor for Detail-Rich StyleGAN Inversion and High Quality Image Editingcited by 0