Research map / Image generation / Editing and style transfer
GAN Inversion for Out-of-Range Images With Geometric Transformations
Kyoungkook Kang, Seongtae Kim, Sunghyun Cho
Read the paperCheck your idea against 137,217 papers
AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.
Measured effects reported in the paper
| What was changed | Where | Measured | Effect |
|---|---|---|---|
| use 16x16 base code instead of 8x8 | BDInvert · CelebA-HQ, Translation 0 | PSNR: 23.69 -> 26.47 | Improved |
| use 16x16 base code instead of 8x8 | BDInvert · CelebA-HQ, Rotation 30 | PSNR: 21.35 -> 26.33 | Improved |
| use finer-scale base code (16x16 vs 8x8) | BDInvert · semantic editing | editing diversity | Hurt |
| use F/W+ space with base code instead of PSP encoder | BDInvert · CelebA-HQ, Translation 0 | PSNR: 20.54 -> 26.47 | Improved |
| apply 30 degree rotation to input images | P-norm+ · CelebA-HQ | PSNR: 3.86 dB | Hurt |
| increase translation from 0 to 150 pixels | Im2StyleGAN · CelebA-HQ | PSNR: 25.63 -> 23.92 | Hurt |
| increase translation from 0 to 150 pixels | BDInvert (16x16) · CelebA-HQ | PSNR: 26.47 -> 26.43 | No clear effect |
| add regularization on detail code w_M+ and base code f (ablation) | BDInvert · CelebA-HQ ablation (qualitative) | semantic editing quality | Improved |
| use BDInvert instead of Im2StyleGAN | BDInvert · LSUN Bedroom | FID: 111.73 -> 49.92 | Improved |
| use BDInvert instead of Im2StyleGAN | BDInvert · LSUN Cat | FID: 71.19 -> 31.74 | Improved |
| use BDInvert instead of Im2StyleGAN | BDInvert · LSUN Tower | PSNR: 20.64 -> 20.37 | Hurt |
Closest papers by meaning
- In-Domain GAN Inversion for Real Image Editing (ECCV 2020 · similarity 0.93)
- In-Domain GAN Inversion for Faithful Reconstruction and Editability (TPAMI 2024 · similarity 0.92)
- Editing Out-of-Domain GAN Inversion via Differential Activations (ECCV 2022 · similarity 0.92)
- GAN Inversion: A Survey (TPAMI 2022 · similarity 0.90)
- Delving StyleGAN Inversion for Image Editing: A Foundation Latent Space Viewpoint (CVPR 2023 · similarity 0.90)
- Gradual Residuals Alignment: A Dual-Stream Framework for GAN Inversion and Image Attribute Editing (AAAI 2024 · similarity 0.89)
Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.
