One Figure, Every Canvas: Editable Flowchart Relayout via Agentic Pipeline
arXiv cs.AI · 6 Oct 2026, 03:59 AEST
A new method for editable flowchart relayout has been developed, addressing the challenges posed by different aspect ratios in machine learning papers. The proposed agentic pipeline significantly improves content fidelity compared to existing methods.
Key facts
- Pipeline 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.
- We formulate aspect-ratio-adaptive flowchart relayout as a distinct task: given a raster flowchart and a target ratio, produce a structurally faithful, hallucination-free, editable layout.
- On a curated benchmark of 100 flowcharts at five aspect ratios, evaluated by Gemini 3.1 Pro and validated against human judgments, our method reaches 68.6% Content Fidelity versus 11.2-41.4% for prior work.
Read aloud
A new method for editable flowchart relayout has been developed, addressing the challenges posed by different aspect ratios in machine learning papers. The proposed agentic pipeline significantly improves content fidelity compared to existing methods. Pipeline figures in ML papers must be repurposed across many canvases, including paper columns, sixteen:nine slides, portrait posters, one:one social teasers, nine:sixteen phone previews. We formulate aspect-ratio-adaptive flowchart relayout as a distinct task: given a raster flowchart and a target ratio, produce a structurally faithful, hallucination-free, editable layout. On a curated benchmark of one hundred flowcharts at five aspect ratios, evaluated by Gemini three point one Pro and validated against human judgments, our method reaches sixty eight point six percent Content Fidelity versus eleven point two-forty one point four percent for prior work.
Sources in this cluster
- One Figure, Every Canvas: Editable Flowchart Relayout via Agentic Pipeline
arXiv cs.AI · 6 Oct 2026, 03:59 AEST