Knowledge Discovery Group

Detecting the Type of an Infographic Using Eye Tracking


Infographics are a common way to visualize research results or data of a study for the human observer. While most information graphics make it easy for an observer to understand the message, it is difficult for computers to process these graphics. Although there are different kinds of classification methods for infographics, none of these focus on the way the target group understands them but focus on either low or high level features.

This work shall analyze the gaze information retrieved from an eye tracking device to give computers a better understanding of how humans understand information graphics. The goal is to automatically categorize these graphics into specific categories (e.g., bar charts, pie charts, ...). Therefore, a user study must be planned, data from an eyetracking device collected and statistically analyzed.

The main aspect are:
- Design of an application for gaze retrieval from infographics
- Implementing a learning-based approach for infographic classification
- Performing an evaluation with a user study


- Good programming skills
- Knowledge of methods for user evaluations and machine learning are of advantage

Got interested? Send an email!


  • Homepage kicked off!