Research map / Image generation / Editing and style transfer
Dotting the Eye: An Intent-Driven Image Retouching Agent for Visual Focus Enhancement
Chujie Qin, Zilong Zhang, Zewei Chang, Chun-Le Guo, Ruixing Wang, Tao Hu, Ming-Ming Cheng, Chongyi Li
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 VLM planner (ablation) | fine-tuned FLUX.1-Kontext baseline · Average over four scenarios | PSNR: 21.70 -> 21.76 | Improved |
| add VLM planner (ablation) | fine-tuned FLUX.1-Kontext baseline · Focus KL | KL: 0.2006 -> 0.2199 | Hurt |
| add VLM planner (ablation) | fine-tuned FLUX.1-Kontext baseline · Global (MIT-FiveK) | PSNR: 18.54 -> 18.07 | Hurt |
| add pseudo-intention attention alignment (PAA) (ablation) | A+Planner · Focus (ControlArt-Bench) | KL: 0.2199 -> 0.1655 | Improved |
| add PAA (ablation) | A+Planner · Local (MMArt-Bench) | PSNR: 27.75 -> 28.55 | Improved |
| add PAA (ablation) | A+Planner · Global (MIT-FiveK) | PSNR: 18.07 -> 17.75 | Hurt |
| add operation-consistency loss (ablation) | B+PAA · Global (MIT-FiveK) | PSNR/SSIM: from 17.75/0.7657 to 23.34/0.800 | Improved |
| add operation-consistency loss (ablation) | B+PAA · Multi-round retouching | PSNR: 20.33 -> 20.81 | Improved |
| add operation-consistency loss (ablation) | B+PAA · Average over four scenarios | PSNR: 22.14 -> 24.17 | Improved |
| add operation-consistency loss (ablation) | B+PAA · Focus (ControlArt-Bench) | KL: 0.1655 -> 0.2215 | Hurt |
| use SAM masks instead of mixed mask types (ablation) | EyeControl · ControlArt-Bench | PSNR: 21.88 -> 21.66 | No clear effect |
| evaluate focus with FA instead of SC | VLM-based metric · Metric validity | winner judgement accuracy: 67.9% -> 87.6% | Improved |
Earlier methods this paper compares with
Named baselines from earlier accepted papers in this paper's results tables: better than 1 of 1 on most shared results.
- JarvisEvo — JarvisEvo: Towards a Self-Evolving Photo Editing Agent with Synergistic Editor-Evaluator Optimization (CVPR 2026) · better on 5 of 7 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
- PerTouch: VLM-Driven Agent for Personalized and Semantic Image Retouching (AAAI 2026 · similarity 0.87)
- JarvisArt: Liberating Human Artistic Creativity via an Intelligent Photo Retouching Agent (NeurIPS 2025 · similarity 0.86)
- RetouchIQ: MLLM Agents for Instruction-Based Image Retouching with Generalist Reward (CVPR 2026 · similarity 0.86)
- Blink: Dynamic Visual Token Resolution for Enhanced Multimodal Understanding (CVPR 2026 · similarity 0.83)
- RetouchFormer: Semi-supervised High-Quality Face Retouching Transformer with Prior-Based Selective Self-Attention (AAAI 2024 · similarity 0.83)
- DreamLight: Towards Harmonious and Consistent Image Relighting (NeurIPS 2025 · similarity 0.83)
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.
