Research map / Image generation / Editing and style transfer
Learning by Planning: Language-Guided Global Image Editing
Check 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 |
|---|---|---|---|
| Use T2ONet with operation planning supervision instead of Pix2pixAug | Language-guided image editing · MA5k-Req test | L1: 0.0928 -> 0.0784 | Helped |
| Use T2ONet with operation planning supervision instead of RL | Language-guided image editing · GIER test | L1: 0.2286 -> 0.0997 | Helped |
| Add final image L1 supervision to operation-sequence supervision (ablation) | T2ONet · MA5k-Req test | L1: 0.0949 -> 0.0784 | Helped |
| Add final image L1 supervision to operation-sequence supervision (ablation) | T2ONet · MA5k-Req test | σ × 10²: 0.0532 -> 0.7190 | Helped |
| Replace final image L1 loss with adversarial loss (ablation) | T2ONet · MA5k-Req test | L1: 0.0784 -> 0.0901 | Hurt |
| Add adversarial loss alongside final image L1 loss (ablation) | T2ONet · MA5k-Req test | L1: 0.0784 -> 0.0801 | Hurt |
| Increase parameter sampling width from h=0 to h=0.01 (ablation) | T2ONet · MA5k-Req test | σ × 10²: 0.7190 -> 1.1008 | Helped |
| Increase parameter sampling width from h=0 to h=0.01 (ablation) | T2ONet · MA5k-Req test | L1: 0.0784 -> 0.0809 | Hurt |
| Increase parameter sampling width from h=0 to h=0.1 (ablation) | T2ONet · MA5k-Req test | σ × 10²: 0.7190 -> 2.1482 | Helped |
| Increase parameter sampling width from h=0 to h=0.1 (ablation) | T2ONet · MA5k-Req test | L1: 0.0784 -> 0.0979 | Hurt |
| Remove intermediate image feedback (ablation) | T2ONet · MA5k-Req test | L1: 0.0784 -> 0.0863 | Hurt |
| Remove previous operation input (ablation) | T2ONet · MA5k-Req test | L1: 0.0784 -> 0.0837 | Hurt |
Closest papers by meaning
- E$^2$GAN: Efficient Training of Efficient GANs for Image-to-Image Translation (ICML 2024 · similarity 0.87)
- FireEdit: Fine-grained Instruction-based Image Editing via Region-aware Vision Language Model (CVPR 2025 · similarity 0.87)
- Meta-CoT: Enhancing Granularity and Generalization in Image Editing (CVPR 2026 · similarity 0.86)
- Guiding Instruction-based Image Editing via Multimodal Large Language Models (ICLR 2024 · similarity 0.85)
- ImgEdit: A Unified Image Editing Dataset and Benchmark (NeurIPS 2025 · similarity 0.85)
- AdapEdit: Spatio-Temporal Guided Adaptive Editing Algorithm for Text-Based Continuity-Sensitive Image Editing (AAAI 2024 · similarity 0.84)
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.