One Figure, Every Canvas: Editable Flowchart Relayout via Agentic Pipeline
An agentic parse-style-layout pipeline relayouts flowcharts across aspect ratios with 68.6% content fidelity.
Researchers define aspect-ratio-adaptive flowchart relayout: given a raster flowchart and a target ratio, produce a structurally faithful, editable layout. Their agentic pipeline separates Parse, Style, and Layout, pairing each agent with a critic that combines deterministic checks and visual-language-model feedback, and emits draw.io-editable mxGraph XML. On 100 flowcharts at five aspect ratios, scored by Gemini 3.1 Pro and checked against human judgments, the method reaches 68.6% Content Fidelity versus 11.2-41.4% for prior work.
- Task: relayout a raster flowchart to a new aspect ratio without breaking edges.
- Pipeline uses Parse, Style, and Layout agents, each paired with a critic.
- Outputs are editable draw.io mxGraph XML rather than a flat image.
- Content Fidelity is 68.6% versus 11.2-41.4% for earlier methods.
- Vendors
- Products
- draw.io
- AI models
- Gemini 3.1 Pro
From the original article
1 min read at arxiv.orgPipeline figures in ML papers must be repurposed across many canvases, including paper columns, 16:9 slides, portrait posters, 1:1 social teasers, 9:16 phone previews. Each format imposes a different aspect ratio on the same computational graph, where any silently broken connection misrepresents the method.
By Shih-Chen Tseng · arxiv.org
A short excerpt; the article and its images belong to the publisher. Summary, key facts and assessment above are AI-written from the article and can be wrong.