Research map / Image generation / Editing and style transfer
Frequency Domain Image Translation: More Photo-Realistic, Better Identity-Preserving
Mu Cai, Hong Zhang, Huijuan Huang, Qichuan Geng, Yixuan Li, Gao Huang
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 |
|---|---|---|---|
| add pixel-space frequency loss to Swap AE (ablation) | Swap AE · LSUN Church | FID: 52.34 -> 49.47 | Improved |
| add Fourier-space frequency loss to Swap AE (ablation) | Swap AE · LSUN Church | FID: 52.34 -> 49.62 | Improved |
| add Fourier-space loss on top of pixel-space loss (ablation) | FDIT · LSUN Church | FID: 49.47 -> 48.21 | Improved |
| add pixel-space loss on top of Fourier-space loss (ablation) | FDIT · LSUN Church | FID: 49.62 -> 48.21 | Improved |
| add frequency losses (FDIT) to Swap AE | Swap AE · Flickr Waterfalls | FID: 50.90 -> 48.76 | Improved |
| add frequency losses (FDIT) to Swap AE | Swap AE · FFHQ | FID: 59.83 -> 55.96 | Improved |
| add frequency losses (FDIT) to Swap AE | Swap AE · CelebA-HQ | FID: 43.47 -> 42.02 | Improved |
| add frequency losses (FDIT) to Swap AE | Swap AE · LSUN Church | FID: 52.34 -> 48.21 | Improved |
| add frequency reconstruction loss to Image2StyleGAN inversion | Image2StyleGAN · GAN inversion 1024x1024 | PSNR: 19.626 -> 20.466 | Improved |
| add frequency reconstruction loss to Image2StyleGAN inversion | Image2StyleGAN · GAN inversion 1024x1024 | MSE: 0.0226 -> 0.0205 | Improved |
| add frequency reconstruction loss to Image2StyleGAN inversion | Image2StyleGAN · GAN inversion 1024x1024 | SSIM: 0.6160 -> 0.6218 | Improved |
| add frequency losses (FDIT) to StarGAN v2 | StarGAN v2 · CelebA-HQ-Smile | FID: 17.32 -> 16.86 | Improved |
Earlier methods this paper compares with
Named baselines from earlier accepted papers in this paper's results tables: better than 2 of 2 on most shared results.
- Image2StyleGAN — Image2StyleGAN++: How to Edit the Embedded Images? (CVPR 2020) · better on 4 of 4 shared results
- StarGAN v2 — StarGAN v2: Diverse Image Synthesis for Multiple Domains (CVPR 2020) · better on 1 of 1 shared result
Matched by method name in results tables read by AI. Authors' own tables tend to favour the newer method, and a comparison can be missed when a method is renamed.
Closest papers by meaning
- Focal Frequency Loss for Image Reconstruction and Synthesis (ICCV 2021 · similarity 0.88)
- Misalignment-Robust Frequency Distribution Loss for Image Transformation (CVPR 2024 · similarity 0.88)
- Frequency-Guided Diffusion for Training-Free Text-Driven Image Translation (ICCV 2025 · similarity 0.88)
- DeCo: Frequency-Decoupled Pixel Diffusion for End-to-End Image Generation (CVPR 2026 · similarity 0.87)
- SEIT: Structural Enhancement for Unsupervised Image Translation in Frequency Domain (AAAI 2024 · similarity 0.87)
- Spectrum Translation for Refinement of Image Generation (STIG) Based on Contrastive Learning and Spectral Filter Profile (AAAI 2024 · similarity 0.86)
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.
