AI Paper Atlas

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

Rethinking Layered Graphic Design Generation with a Top-Down Approach

Jingye Chen, Zhaowen Wang, Nanxuan Zhao, Li Zhang, Difan Liu, Jimei Yang, Qifeng Chen

ICCV 2025 · Image generation › Text-to-image diffusion · approach: Layout-to-image control · 9 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 questionnaire-based result selection (ablation)Accordion text removal · Text removalPSNR: 18.01 -> 21.18Improved
add questionnaire-based result selection (ablation)Accordion object removal · Object removalPSNR: 22.92 -> 24.87Improved
joint multi-task training of the VLM instead of separate (ablation)Accordion VLM · Design39Kaverage score: 72.03 -> 72.85No clear effect
joint training instead of separate (ablation)Accordion VLM · Design39K original designsText Recognition NED: 72.87 -> 68.51Hurt
joint training instead of separate (ablation)Accordion VLM · Design39K original designsFont Accuracy: 24.51 -> 21.62Hurt
joint training instead of separate (ablation)Accordion VLM · Design39K backgrounds without textObject Detection F1: 83.52 -> 86.94Improved
joint training instead of separate (ablation)Accordion VLM · Design39K nonsensical text designsObject Detection F1: 79.27 -> 83.06Improved
add OCR prompt to the system (ablation)Accordion VLM · Design39K original designsparagraph-level OCR NED: 61.28 -> 68.51Improved
add OCR prompt to the system (ablation)Accordion VLM · Design39Ktext detection F1: 73.12 -> 78.59Improved
refine Flux reference images with AccordionFlux · DesignIntentioncontent relevance (GPT-4V): 7.0 -> 7.4Improved
top-down Accordion instead of bottom-up COLECOLE · DesignIntentionGPT-4V average score: 6.0 -> 6.5Improved
Accordion instead of COLE for aestheticsCOLE · DesignIntentionLAION aesthetic score: 4.72 -> 4.98Improved

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