Research map / Image generation / Editing and style transfer
Eta Inversion: Designing an Optimal Eta Function for Diffusion-based Real Image Editing
Wonjun Kang, Kevin Galim, Hyung Il Koo
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 Eta Inversion (injecting real noise via a designed eta in DDIM sampling) instead of Direct Inversion | PtP on PIE-Bench · PIE-Bench, PtP | CLIP similarity: 0.3091879120469093 -> 0.31247970819473 | Improved |
| use Eta Inversion (2) instead of Direct Inversion | PnP on PIE-Bench · PIE-Bench, PnP | CLIP similarity: 0.3131952231909548 -> 0.3162599997861045 | Improved |
| use region-dependent (masked) eta instead of global eta (ablation) | Eta Inversion (1) · PIE-Bench, PtP | BG-LPIPS: 0.06737695838071626 -> 0.06568354647575429 | Improved |
| use region-dependent (masked) eta instead of global eta (ablation) | Eta Inversion (2) · PIE-Bench, PtP | CLIP similarity: 0.3126584818959236 -> 0.31247970819473 CLIP; DINO 0.018496008312795312 -> 0.017007591246760316 | Mixed |
| increase eta (EtaInv (2) vs (1)) to promote more editing (ablation) | Eta Inversion · PIE-Bench, PtP | CLIP similarity vs structural distance: CLIP 0.31007335275411 -> 0.31247970819473; DINO 0.013433658024296165 -> 0.017007591246760316 | Mixed |
| use larger eta without masking for style transfer | Eta Inversion (3) · PIE-Bench change-style, PtP | CLIP similarity: 0.31713469438254 -> 0.3285034172236919 | Improved |
| use larger eta for style transfer | Eta Inversion (3) · PIE-Bench change-style, PtP | DINO structural distance: 0.016382710146717726 -> 0.04185123109491542 | Hurt |
| use a decreasing linear eta schedule (ablation) | Eta Inversion · PIE-Bench | editing quality | Improved |
Closest papers by meaning
- Inversion-Free Image Editing with Language-Guided Diffusion Models (CVPR 2024 · similarity 0.92)
- PnP Inversion: Boosting Diffusion-based Editing with 3 Lines of Code (ICLR 2024 · similarity 0.92)
- Object-Aware Inversion and Reassembly for Image Editing (ICLR 2024 · similarity 0.91)
- Lightning-Fast Image Inversion and Editing for Text-to-Image Diffusion Models (ICLR 2025 · similarity 0.91)
- Noise Map Guidance: Inversion with Spatial Context for Real Image Editing (ICLR 2024 · similarity 0.90)
- DCI: Dual-Conditional Inversion for Boosting Diffusion-Based Image Editing (NeurIPS 2025 · similarity 0.90)
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.
