AI Paper Atlas

Research map / Vision-language and multimodal / Human and embodied modeling

FlowInOne: Unifying Multimodal Generation as Image-in, Image-out Flow Matching

Junchao Yi, Rui Zhao, Jiahao Tang, Weixian Lei, Linjie Li, Qisheng Su, Zhengyuan Yang, Lijuan Wang, Xiaofeng Zhu, Alex Jinpeng Wang

ECCV 2026 · Vision-language and multimodal › Human and embodied modeling · approach: Vision-language pretraining · 2 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
train jointly on all 5M samples instead of two-stage (ablation)FlowInOne · VP-Benchavg pass rate: 29.1% -> 47.8%Improved
add dual-path cross-attention to source latent (ablation)FlowInOne · VP-Benchavg pass rate: 0.182 -> 0.214Improved
add spatially-adaptive gating to dual-path cross-attention (ablation)FlowInOne · VP-Benchavg pass rate: 0.214 -> 0.231Improved
use MLP+MLP compression instead of truncation (ablation)FlowInOne · VP-Benchavg pass rate: 0.169 -> 0.182Improved
use MLP+MLP compression instead of VAE expansion (ablation)FlowInOne · VP-Benchavg pass rate: 0.157 -> 0.182Improved
render instructions into image canvas (image-in image-out) vs text-conditioned open-source editorsFlowInOne · VP-Bench force understanding (Gemini3)success rate: .727 vs .047 (Qwen-IE-2509)Improved
FlowInOne vs commercial Nano Banana on total successFlowInOne · VP-Benchtotal success rate: .540 vs .589 (Gemini3); .392 vs .391 (GPT5.2)Mixed
FlowInOne vs Nano Banana on visual realismFlowInOne · VP-Bench (Gemini3)visual realism: 4.43 -> 3.12Hurt
remove in-image text, keep visual prompts only (ablation)FlowInOne · visual instruction ablation (4 samples)qualitativeHurt
replace valid instruction text with random corrupted text (ablation)FlowInOne · robustness testqualitativeHurt
lower input resolution to 128x128 (ablation)FlowInOne · robustness testqualitativeHurt
change text style/length/blur of rendered instruction (ablation)FlowInOne · robustness testqualitativeNo clear effect

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.

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