Research map / Image generation / Text-to-image diffusion
GIDE: Unlocking Diffusion LLMs for Precise Training-Free Image Editing
Zifeng Zhu, Jiaming Han, Jiaxiang Zhao, Minnan Luo, Xiangyu Yue
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 discrete noise inversion (ablation) | GIDE · GIDE-Bench (ablation) | PSNR: 17.85 -> 20.40 | Improved |
| add spatial grounding (ablation) | GIDE · GIDE-Bench (ablation) | PSNR: 9.42 -> 20.40 | Improved |
| add spatial grounding (ablation) | GIDE · GIDE-Bench (ablation) | SC (Gemini): 2.68 -> 4.26 | Improved |
| add visual refinement (ablation) | GIDE · GIDE-Bench (ablation) | PSNR: 17.23 -> 20.40 | Improved |
| add intrinsic refinement (ablation) | GIDE · GIDE-Bench (ablation) | SC (GPT): 3.96 -> 4.47 | Improved |
| add intrinsic refinement (ablation) | GIDE · GIDE-Bench (ablation) | PSNR: 19.80 -> 20.40 | Improved |
| add residual recovery (ablation) | GIDE · GIDE-Bench (ablation) | PSNR: 17.65 -> 20.40 | Improved |
| add discrete inversion (ablation) | GIDE · GIDE-Bench (ablation) | PQ (Gemini): 3.28 -> 3.78 | Improved |
| mix inversion residual with lambda 0.2 instead of 0.0 (ablation) | GIDE · GIDE-Bench (sensitivity) | SC (GPT): 4.30 -> 4.47 | Improved |
| increase lambda from 0.2 to 0.4 (ablation) | GIDE · GIDE-Bench (sensitivity) | SC (GPT): 4.47 -> 4.21 | Hurt |
| increase lambda from 0.2 to 0.1 (ablation) | GIDE · GIDE-Bench (sensitivity) | MSE: 1224.89 -> 1306.47 | No clear effect |
| apply GIDE to Lumina-DiMOO | Lumina-DiMOO · GIDE-Bench | SC (GPT): 3.10 -> 4.47 | Improved |
Closest papers by meaning
- Inversion-Free Image Editing with Language-Guided Diffusion Models (CVPR 2024 · similarity 0.92)
- EditInfinity: Image Editing with Binary-Quantized Generative Models (NeurIPS 2025 · similarity 0.90)
- Schedule Your Edit: A Simple yet Effective Diffusion Noise Schedule for Image Editing (NeurIPS 2024 · similarity 0.90)
- Invertible Consistency Distillation for Text-Guided Image Editing in Around 7 Steps (NeurIPS 2024 · similarity 0.89)
- Enabling Instructional Image Editing with In-Context Generation in Large Scale Diffusion Transformer (NeurIPS 2025 · similarity 0.89)
- Drag Your Noise: Interactive Point-based Editing via Diffusion Semantic Propagation (CVPR 2024 · similarity 0.89)
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.
