AI Paper Atlas

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

VisualCloze: A Universal Image Generation Framework via Visual In-Context Learning

Zhong-Yu Li, Ruoyi Du, Juncheng Yan, Le Zhuo, Zhen Li, Peng Gao, Zhanyu Ma, Ming-Ming Cheng

ICCV 2025 · Image generation › Text-to-image diffusion · approach: Layout-to-image control · 33 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
fine-tune FLUX.1-Fill-dev instead of FLUX.1-dev (ablation)VisualCloze · Depth-to-imageRMSE: 25.06 -> 10.31Improved
fine-tune FLUX.1-Fill-dev instead of FLUX.1-dev (ablation)VisualCloze · Canny-to-imageF1: 0.39 -> 0.35Hurt
fine-tune FLUX.1-Fill-dev instead of FLUX.1-dev (ablation)VisualCloze · Subject-driven generationDINOv2: 78.05 -> 80.41Improved
fine-tune FLUX.1-Fill-dev instead of FLUX.1-dev (ablation)VisualCloze · Style transferCLIP text alignment: 0.30 -> 0.29No clear effect
add two in-context examples (ablation)VisualCloze (Fill) · Depth-to-imageRMSE: 10.31 -> 9.68Improved
add two in-context examples (ablation)VisualCloze (Fill) · DeblurringRMSE: 26.53 -> 25.57Improved
add two in-context examples (ablation)VisualCloze (Fill) · Canny-to-imageFID: 30.60 -> 31.15Hurt
add two in-context examples (ablation)VisualCloze (Fill) · Subject-driven generationDINOv2: 80.41 -> 80.32No clear effect
add one in-context example (ablation)VisualCloze (Fill) · DeblurringFID: 40.59 -> 36.93Improved
use VisualCloze (Fill) instead of OmniGenVisualCloze · Style transferCLIP text alignment: 0.27 -> 0.29Improved
use VisualCloze (Fill) instead of OmniGenVisualCloze · Canny-to-imageFID: 51.58 -> 30.60Improved
use VisualCloze (Fill) instead of OminiControl specialistVisualCloze · Subject-driven generationDINOv2: 73.17 -> 80.41Improved

Earlier methods this paper compares with

Named baselines from earlier accepted papers in this paper's results tables: better than 1 of 2 on most 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

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