Research map / Image generation / Text-to-image diffusion
Frequency-Guided Diffusion for Training-Free Text-Driven Image Translation
Zheng Gao, Jifei Song, Zhensong Zhang, Jiankang Deng, Ioannis Patras
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 low-frequency substitution (ablation) | DDPM-Edit baseline · Aesthetics style-guided | CSD: 0.530 -> 0.631 | Improved |
| add color alignment (AdaIN) on top of low-frequency substitution (ablation) | FGD low-frequency guidance · Aesthetics style-guided | CSD: 0.631 -> 0.649 | Improved |
| add color alignment alone (ablation) | DDPM-Edit baseline · Aesthetics style-guided | CSD: 0.530 -> 0.539 | Improved |
| add high-frequency substitution of latents (ablation) | DDPM-Edit baseline · Aesthetics structure-guided | DINO distance: 0.051 -> 0.045 | Improved |
| add high-frequency substitution of latents (ablation) | DDPM-Edit baseline · Aesthetics structure-guided | CLIP score: 0.269 -> 0.265 | Hurt |
| add high-frequency substitution of latents (ablation) | DDPM-Edit baseline · Aesthetics structure-guided | HPSv2: 27.28 -> 27.03 | Hurt |
| add high-frequency injection (ablation) | DDPM-Edit baseline · Aesthetics structure-guided | DINO distance: 0.051 -> 0.037 | Improved |
| add high-frequency alignment (ablation) | DDPM-Edit baseline · Aesthetics structure-guided | Edge similarity: 0.910 -> 0.927 | Improved |
| add high-frequency alignment to injection (ablation) | FGD high-frequency guidance · Aesthetics structure-guided | Edge similarity: 0.915 -> 0.931 | Improved |
| add high-frequency injection and alignment instead of substitution (ablation) | FBSDiff-style baseline · Aesthetics structure-guided | HPSv2: 27.03 -> 27.56 | Improved |
| replace FBSDiff with FGD | FBSDiff · Aesthetics structure-guided | DINO distance: 0.058 -> 0.035 | Improved |
| replace FBSDiff with FGD | FBSDiff · Aesthetics style-guided | CSD: 0.436 -> 0.649 | Improved |
Earlier methods this paper compares with
Named baselines from earlier accepted papers in this paper's results tables: better than 1 of 2 on most shared results.
- LEDITS++ — LEDITS++: Limitless Image Editing using Text-to-Image Models (CVPR 2024) · better on 0 of 2 shared results
- FreeControl — FreeControl: Training-Free Spatial Control of Any Text-to-Image Diffusion Model with Any Condition (CVPR 2024) · 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
- Frequency-Controlled Diffusion Model for Versatile Text-Guided Image-to-Image Translation (AAAI 2024 · similarity 0.90)
- Diffusion-based Image Translation using disentangled style and content representation (ICLR 2023 · similarity 0.90)
- One Stone with Two Birds: A Null-Text-Null Frequency-Aware Diffusion Models for Text-Guided Image Inpainting (NeurIPS 2025 · similarity 0.89)
- Inversion-Free Image Editing with Language-Guided Diffusion Models (CVPR 2024 · similarity 0.89)
- Noise Diffusion for Enhancing Semantic Faithfulness in Text-to-Image Synthesis (CVPR 2025 · similarity 0.89)
- Disentangling Dual Image References in Frequency Aware Diffusion Models for Personalized Generation (NeurIPS 2026 · similarity 0.88)
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.
