Turn an academic paper (PDF, DOI, URL, pasted text, or extracted manuscript) into a composed vertical academic poster image using a divide-and-compose workflow. Use when the user asks for “学术海报”, “论文海报”, “conference poster”, “research poster”, “A0 poster”, “把论文做成海报”, “灵魂海报 / 设计感海报 / 特色锚点海报 / soulful poster”, or a visually attractive poster summarizing a paper. The workflow reads and understands the paper first, decomposes the poster into semantic regions and visual assets, generates/redraws sub-images where useful, preserves evidence crops when source fidelity matters, then deterministically composes the final poster with local text, layout, and QC.
2026-05-08