FigureOUT: A Personalized Tool for Summarizing and Visualizing Academic Literature

A class project from KAIST ID503 Design Project (Prof. Tak Yeon Lee), by Team 4F. I worked on ideation and front-end development.

Problem — finding the right paper is hard

Whether you’re a novice or an expert, finding the paper is a slog. Search engines return only thin, text-based metadata — title, authors, abstract — which rarely says enough to judge a paper at a glance. And everyone sees the same summary, regardless of their purpose or interest. We wanted a personalized way to summarize and visualize papers so researchers can quickly tell whether something is worth their time.

Solution — visual, personalized paper summaries

FigureOUT summarizes papers visually, in four ways:

• Figure Cards — search results as cards with thumbnails pulled from each paper.
• Proportion Bar — shows how much space each section takes, so you can guess a paper’s structure without opening the PDF.
• Content Icons — label a paper by its number of graphs, images, and formulas.
• Figure Wall — gathers images from many papers into one frame to compare them and explore vague ideas.

Thumbnail extraction is personalized to each researcher’s real-time search history and stated interests, and a filter lets users switch between an AI-driven and a self-chosen visual summary.

Value

By making relevance visible, FigureOUT saves the time researchers spend hunting and triaging papers — time that can go back into the research itself.