AI Paper Atlas

Research map / Image generation / Fast diffusion sampling

Region-Aware Test-Time Scaling for Compositional Image Generation

Mingzhu Shen, Peng Ye, Xinyin Ma, Gongfan Fang, Christos-Savvas Bouganis, Yiren Zhao, Xinchao Wang

ECCV 2026 · Image generation › Fast diffusion sampling · approach: Distillation for few-step generation · 0 citations (OpenAlex)

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 RAS (region-aware generation) at a single sampleFLUX.1-dev · GenEvaloverall score: 0.67 -> 0.77Improved
add RAS instead of noise scaling at 32 samplesFLUX.1-dev · GenEvaloverall score: 0.85 (N=32 noise) vs 0.85 (RAS N=2)Improved
add RAS with 4 samples versus prompt scaling with 32FLUX.1-dev · GenEvaloverall score: 0.87 -> 0.88Improved
add RAS to ReflectionFlowReflectionFlow · GenEvaloverall score: 0.91 -> 0.92Improved
add RAS to SANA-1.0-1.6B with 8 samples versus 20 noise samplesSANA-1.0-1.6B · GenEvaloverall score: 0.80 -> 0.81Improved
add regional guidance (RAS) to FLUX at N=1FLUX.1-dev · GenEval Positionaccuracy: 0.21 -> 0.67Improved
add LLM rewrite to vanilla FLUX (ablation)FLUX.1-dev · GenEval Positionaccuracy: 0.21 -> 0.40Improved
add LLM planning and RAS after rewrite (ablation)FLUX.1-dev · GenEval Positionaccuracy: 0.40 -> 0.60Improved
add Best-of-2 selection on top of RAS (ablation)FLUX.1-dev · GenEval Positionaccuracy: 0.60 -> 0.77Improved
add a single region prompt to noise-only scaling (ablation)FLUX.1-dev · GenEval Positionprompt-level success rate: 82% -> 89%Improved
balance budget across noise/global/region (2/4/4) versus all noise (ablation)FLUX.1-dev · GenEval Positionprompt-level success rate: 82% -> 92%Improved
use GPT-4o instead of Qwen2-VL-7B as planner (ablation)RAS · GenEval Positionlayout parsing accuracy: 52% -> 90%Improved

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