Research map / Image generation / Editing and style transfer
StyleGANEX: StyleGAN-Based Manipulation Beyond Cropped Aligned Faces
Shuai Yang, Liming Jiang, Ziwei Liu, Chen Change Loy
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 Step I encoder prediction before Step II optimization in two-step inversion (ablation) | StyleGANEX inversion · Normal FoV face inversion | LPIPS: 0.120 -> 0.086 | Improved |
| add Step II latent and feature optimization after encoder (ablation) | StyleGANEX inversion · Normal FoV face inversion | LPIPS: 0.385 -> 0.086 | Improved |
| extract w+ from cropped aligned face instead of whole image (ablation) | StyleGANEX encoder · inversion (qualitative) | restoration quality | Improved |
| add skip connections (l>0) from encoder to generator (ablation) | StyleGANEX translation · sketch/mask-to-face (qualitative) | detail preservation | Mixed |
| use StyleGANEX instead of StyleGAN refactor, dilated shallow layers | StyleGAN · image generation time | seconds per image: 0.026s -> 0.028s | No clear effect |
| use StyleGANEX vs VToonify-T for video toonification | video toonification · user study | preference score: 0.083 -> 0.917 | Improved |
| use StyleGANEX vs pix2pixHD for sketch-to-face | sketch-to-face · user study | preference score: 0.400 -> 0.450 | Improved |
| StyleGANEX vs HyperStyle for video face editing | video face editing · FaceForensics++ videos | FVD: 432.64 -> 198.35 | Improved |
Earlier methods this paper compares with
Named baselines from earlier accepted papers in this paper's results tables: better than 1 of 1 on most shared results.
- HyperStyle — HyperStyle: StyleGAN Inversion With HyperNetworks for Real Image Editing (CVPR 2022) · better on 4 of 4 shared results
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
- Face Generation and Editing With StyleGAN: A Survey (TPAMI 2024 · similarity 0.91)
- StyleGAN2 Distillation for Feed-forward Image Manipulation (ECCV 2020 · similarity 0.90)
- StyleIPSB: Identity-Preserving Semantic Basis of StyleGAN for High Fidelity Face Swapping (CVPR 2023 · similarity 0.89)
- MaskGAN: Towards Diverse and Interactive Facial Image Manipulation (CVPR 2020 · similarity 0.88)
- The Devil is in the Details: StyleFeatureEditor for Detail-Rich StyleGAN Inversion and High Quality Image Editing (CVPR 2024 · similarity 0.87)
- Enhancing Identity-Deformation Disentanglement in StyleGAN for One-Shot Face Video Re-Enactment (AAAI 2025 · 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.
