Paper2Poster - An academic poster generation framework launched by the Universities of Waterloo, National University of Singapore, and Oxford.

What is Paper2Poster?

Paper2Poster is an innovative academic framework developed by institutions such as the University of Waterloo in Canada and the National University of Singapore. It generates posters from scientific papers using multimodal automation technology. Paper2Poster introduces PosterAgent, a top-down multi-agent system that supports the compression and transformation of long academic papers into structured visual posters. The system achieves efficient generation based on a parser, planner, and painter-commenter loop. Paper2Poster incorporates the PaperQuiz evaluation method, simulating reader responses to measure the poster's ability to convey core content. Paper2Poster excels in visual quality and textual coherence, significantly improving generation efficiency and providing an efficient and low-cost solution for academic poster production.

Main functions of Paper2Poster

• Long text compressionCompress multi-page scientific papers into single-page posters while retaining the core content of the papers.
• Multimodal content processingExtract multimodal content such as text, charts, and images from the paper and integrate it into the poster in a reasonable way.
• Layout optimizationGenerate aesthetically pleasing and logically clear poster layouts, ensuring that content is reasonably distributed within a limited space.
• Visual quality improvement: Optimize the visual effects of posters based on visual feedback mechanisms to improve their readability and attractiveness.
• Evaluation and optimizationThe ability of posters to convey core content is evaluated based on PaperQuiz, and optimizations are made based on feedback.

The technical principles of Paper2Poster

• ParserUse tools (such as MARKER and DOCLING) to convert PDFs to Markdown format and generate a structured JSON-formatted asset library based on LLM.
• PlannerAlign text and visual elements in the parser-generated asset library to generate a binary tree layout. Based on the binary tree layout strategy, estimate the size of each panel according to the content length, maintaining reading order and spatial balance. Use LLM for semantic matching to align each visual element with the most relevant text portion.
• Painter – CommenterThe process generates the specific content for each panel and optimizes the panel layout based on visual feedback. The plotter aligns text and images and generates executable code to render the panels using the python-pptx library. The commenter is a VLM that provides feedback based on zoom-in reference prompts to ensure that panel content does not overflow and is laid out reasonably.

Paper2Poster project address

• Project official website:https://paper2poster.github.io/
• GitHub repository:https://github.com/Paper2Poster/Paper2Poster
• HuggingFace model library:https://huggingface.co/datasets/Paper2Poster/Paper2Poster
• arXiv technical paper:https://arxiv.org/pdf/2505.21497

Application scenarios of Paper2Poster

• Academic ConferenceIt helps researchers quickly convert their papers into posters for conference presentations, saving time and effort.
• Academic ReportThe generated posters serve as supplementary material for academic reports, helping audiences better understand the research content.
• Scientific Research Achievements ShowcaseIn research institutions or laboratories, it serves to showcase the latest research findings, facilitating exchange and learning among peers.
• EducationTeachers can generate posters for teaching purposes to help students understand complex academic concepts more intuitively.
• Online academic platformProvide automated poster generation tools for online academic platforms to enhance user experience and promote academic exchange.