AI Paper Atlas

Research map / Image generation / Text-to-image diffusion

ImageDoctor: Diagnosing Text-to-Image Generation via Grounded Image Reasoning

Yuxiang Guo, Jiang Liu, Ze Wang, Hao Chen, Ximeng Sun, Yang Zhao, Jialian Wu, Xiaodong Yu, Zicheng Liu, Emad Barsoum

ICLR 2026 · Image generation › Text-to-image diffusion · approach: T2I evaluation and controllable adapters

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 changedWhereMeasuredEffect
add task tokens (<ART>, <MIS>) to the heatmap decoder (ablation)ImageDoctor heatmap decoder · RichHF-18Kheatmap CC (artifact / misalignment): 0.508 -> 0.532 (artifact CC); 0.123 -> 0.165 (misalignment CC)Improved
add the look step (flaw bounding boxes) (ablation)ImageDoctor cold start stage 2 · RichHF-18Kmisalignment CC: 0.160 -> 0.224Improved
add the think step (reasoning) (ablation)ImageDoctor cold start stage 2 · RichHF-18Kaverage PLCC: 0.708 -> 0.720Improved
add the grounding reward in reinforcement finetuning (ablation)ImageDoctor RFT · RichHF-18Kaverage PLCC: 0.734 -> 0.741Improved
add reinforcement finetuning after cold start stage 2 (ablation)ImageDoctor · RichHF-18Kaverage PLCC: 0.720 -> 0.741Improved
add heatmap prediction to cold start stage 1 score training (ablation)ImageDoctor cold start stage 1 · RichHF-18Kaverage PLCC: 0.660 -> 0.655No clear effect
run ImageDoctor without the reasoning chain at inference (ablation)ImageDoctor · RichHF-18Kaverage PLCC: 0.741 -> 0.729Hurt
use ImageDoctor scores instead of PickScore as the Flow-GRPO rewardFlow-GRPO on SD3.5-medium · DrawBenchImageReward: 1.002 -> 1.029Improved
use ImageDoctor scores instead of RichHF as the Flow-GRPO rewardFlow-GRPO on SD3.5-medium · DrawBenchImageReward: 0.879 -> 1.029Improved
add heatmap-based dense pixel-level reward (DenseFlow-GRPO) to scalar ImageDoctor rewardFlow-GRPO with ImageDoctor · DrawBenchImageReward: 1.029 -> 1.100Improved
use PickScore as Flow-GRPO rewardFlow-GRPO on SD3.5-medium · DrawBenchCLIPScore: 0.951 -> 0.941Hurt

Closest papers by meaning

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.

Open scime draft

Get the next scime Dispatch and a monthly field update

Free. We send a confirmation link first; every email has a one-click unsubscribe. Privacy