Research map / Image generation / GANs
High Fidelity GAN Inversion via Prior Multi-Subspace Feature Composition
Read the paperCheck your idea against 108,967 papers
AI-extracted summary. Written by AI from the paper's title only (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 |
|---|---|---|---|
| Compose features separately in discovered subspaces instead of across all channels | mGANprior inversion · CelebA | LPIPS: 0.4432 -> 0.4068 | Helped |
| Compose features separately in discovered subspaces instead of across all channels | mGANprior inversion · CelebA | PSNR: 23.59 -> 27.88 | Helped |
| Compose features separately in discovered subspaces instead of across all channels | mGANprior inversion · Bedroom | LPIPS: 0.1578 -> 0.1300 | Helped |
| Compose features separately in discovered subspaces instead of across all channels | mGANprior inversion · Bedroom | PSNR: 25.13 -> 27.59 | Helped |
| Compose features separately in discovered subspaces instead of across all channels | mGANprior inversion · Church | LPIPS: 0.1799 -> 0.1501 | Helped |
| Compose features separately in discovered subspaces instead of across all channels | mGANprior inversion · Church | PSNR: 22.76 -> 29.67 | Helped |
| Use prior subspaces for feature composition | mGANprior colorization · CelebA | AuC: 90.35* -> 91.50 | Helped |
| Use prior subspaces for feature composition | mGANprior colorization · Bedroom | AuC: 90.02 -> 91.34 | Helped |
| Use prior subspaces for feature composition | mGANprior colorization · Church | AuC: 89.43 -> 90.44 | Helped |
| Use prior subspaces for feature composition | mGANprior super-resolution · Face super-resolution (16×) | PSNR: 26.93 -> 28.90 | Helped |
| Use prior subspaces for feature composition | mGANprior super-resolution · Face super-resolution (16×) | LPIPS: 0.3584 -> 0.3450 | Helped |
| Use prior subspaces for feature composition | mGANprior super-resolution · Face super-resolution (16×) | NIQE: 3.19 -> 2.50 | Helped |
Closest papers by meaning
- High-Fidelity 3D GAN Inversion by Pseudo-Multi-View Optimization (CVPR 2023 · similarity 0.91)
- Spatially-Adaptive Multilayer Selection for GAN Inversion and Editing (CVPR 2022 · similarity 0.90)
- High-Fidelity GAN Inversion for Image Attribute Editing (CVPR 2022 · similarity 0.89)
- WarpGAN: Warping-Guided 3D GAN Inversion with Style-Based Novel View Inpainting (NeurIPS 2025 · similarity 0.88)
- Gradual Residuals Alignment: A Dual-Stream Framework for GAN Inversion and Image Attribute Editing (AAAI 2024 · similarity 0.87)
- Spatial-Contextual Discrepancy Information Compensation for GAN Inversion (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.