Research map / Image generation / Editing and style transfer
CogniEdit: Dense Gradient Flow Optimization for Fine-Grained Image Editing
Yan Li, Lin Liu, Xiaopeng Zhang, Wei Xue, Wenhan Luo, Yike Guo, Qi Tian
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 Dense GRPO optimization (ablation) | CogniEdit (Qwen-Image-Edit) · KRIS-Bench Knowledge Reasoning | Avg score: 46.11 -> 59.29 | Improved |
| add Dynamic Token Focus Relocation (ablation) | CogniEdit (Qwen-Image-Edit) · KRIS-Bench Knowledge Reasoning | Avg score: 54.66 -> 59.29 | Improved |
| add both components to Qwen-Image-Edit-r1 (ablation) | Qwen-Image-Edit-r1 · KRIS-Bench Knowledge Reasoning | Avg score: 55.48 -> 59.29 | Improved |
| add Dynamic Token Focus Relocation without Dense GRPO (ablation) | Qwen-Image-Edit-r1 · KRIS-Bench Knowledge Reasoning | Avg score: 55.48 -> 46.11 | Hurt |
| add Dense GRPO without Dynamic Token Focus Relocation (ablation) | Qwen-Image-Edit-r1 · KRIS-Bench Knowledge Reasoning | Avg score: 55.48 -> 54.66 | No clear effect |
| add Dense GRPO and dynamic focus to Qwen-Image-Edit-r1 (ablation) | Qwen-Image-Edit-r1 · KRIS-Bench Knowledge Reasoning | IF: 35.03 -> 40.14 | Improved |
| add VLM instruction reasoning (r1) to Qwen-Image-Edit | Qwen-Image-Edit · KRIS-Bench Social Science | Avg score: 66.40 -> 77.99 | Improved |
| train CogniEdit on top of Qwen-Image-Edit-r1 | Qwen-Image-Edit-r1 · KRIS-Bench Social Science | Avg score: 77.99 -> 77.64 | No clear effect |
| train CogniEdit on top of Qwen-Image-Edit-r1 | Qwen-Image-Edit-r1 · KRIS-Bench Natural Science | Avg score: 64.78 -> 67.42 | Improved |
| Dense GRPO versus standard GRPO (ablation) | CogniEdit · CLIP score vs FLOPs | CLIP score | Improved |
Earlier methods this paper compares with
Named baselines from earlier accepted papers in this paper's results tables: better than 2 of 2 on most shared results.
- AnyEdit — AnyEdit: Mastering Unified High-Quality Image Editing for Any Idea (CVPR 2025) · better on 3 of 3 shared results
- InstructPix2Pix — InstructPix2Pix: Learning To Follow Image Editing Instructions (CVPR 2023) · better on 3 of 3 shared results
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
- AutoEdit: Automatic Hyperparameter Tuning for Image Editing (NeurIPS 2025 · similarity 0.89)
- AnyEdit: Mastering Unified High-Quality Image Editing for Any Idea (CVPR 2025 · similarity 0.89)
- FireEdit: Fine-grained Instruction-based Image Editing via Region-aware Vision Language Model (CVPR 2025 · similarity 0.89)
- InsightEdit: Towards Better Instruction Following for Image Editing (CVPR 2025 · similarity 0.89)
- Enabling Instructional Image Editing with In-Context Generation in Large Scale Diffusion Transformer (NeurIPS 2025 · similarity 0.88)
- Meta-CoT: Enhancing Granularity and Generalization in Image Editing (CVPR 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.
